Self quantification is the trend that just keeps on going. There’s an ever-expanding world of wellness wearables and fitness trackers targeting consumers with shiny promises of the personal value to be had if they monitor stuff like their heart rate, activity and sleep — from smart watches, bands and rings, to smart scales, CGMs (continuous […]

© 2023 TechCrunch. All rights reserved. For personal use only.

Indian fitness platform, Ultrahuman, is expanding its wearable portfolio by launching a smart ring to boost its ability to provide tech loving ‘biohackers’ — and, it hopes, health-concerned Boomers — with more insightful metabolic insights.

Sensors embedded in the forthcoming Ultrahuman Ring include temperature, heart rate and movement monitors, which enable the device to track the wearer’s sleep quality, stress levels and activity density, per CEO and co-founder, Mohit Kuma.

The device is designed to work in conjunction with the startup’s existing wearable, a continuous glucose monitor (CGM) sensor-based service it brands ‘Cyborg’, to deepen the quality of insights for users — such as by identifying when a poor glucose response might be linked to a bad night’s sleep, say, or elevated stress levels, rather than putting all the focus on whatever it was the user ate right before their blood sugar spiked.

The Ultrahuman Ring is not a CGM itself but it can function as a standalone health tracker, according to Kumar — giving the fitness startup a shot at broadening the appeal of its metabolic tracking service since the smart ring just slips on the finger, instead of, as is the case with the CGM, requiring that a spring-loaded filament is fired into the user’s upper arm (and left in place, ‘worn’ under the skin).

The clean, chunky look of the ring band (which comes in a shiny metal titanium or black finish) is also more likely to fit in with fashion-conscious consumers than rocking a ‘Cyborg’ arm patch. 

The Ultrahuman Ring goes up for pre-order today, with shipping slated to start in August.

At the time of writing pricing hasn’t been confirmed but the startup told us there will be two options: One (premium) price covering lifetime usage; and another (monthly) subscription option with a relatively small lock in period, after which the user would be free to quit on demand.

One ring to end the guesswork?

“The idea is to help you understand more about what are the additional factors in your metabolism,” say Kumar, discussing the incoming smart ring in a Zoom call with TechCrunch. “Right now today with the glucose monitor you actually understand the outcome of how glucose metabolism works but there are many other factors that affect glucose levels — factors like stress, sleep, activity. These are the major ones.”

“Today, a lot of this is actually guesswork,” he goes on. “But with our own wearable — and with the access to the raw data of the wearable — we actually have the ability now to understand what was the leading factor that led to a poorer glucose response. For example, if you’re under-recovered because of, let’s say, lack of sleep and the glucose levels get elevated the platform now can clearly figure out what is the contributing factor. And similarly for lack of activity.”

Many factors can affect how the body metabolises glucose, while big swings in blood sugar can be associated with health problems like diabetes and heart disease — creating an impetus for consumers to make lifestyle changes intended to stablize their glucose response, such as upping their activity level, choosing a healthier diet and getting enough sleep.

Ultrahuman’s metabolic fitness tracking service essentially sells real-time feedback to help individuals get a handle on what’s going on with their biology. But when we road-tested its beta product last year, we highlighted the relative challenge for the average user to intelligently interpret their glucose variability data — and link it to specific lifestyle factors — vs taking a too simplistic read of the data.

The smart ring looks intended to shrink this interpretation gap by enabling Ultrahuman’s platform to track and triangulate a variety of biomarkers to provide the user with a stronger read on what’s behind their glucose peaks and troughs. (Or: “If lack of activity is leading to elevated base levels of glucose the platform will be able to decode it in a much more efficient way,” as Kumar puts it.)

Ultrahuman will be going up against a number of more established players in the smart ring space — typically also with a strong focus on health/fitness tracking. However it argues its differentiating twist here is that it’s “optimizing for metabolism” — and, well, it has the glucose tracking data to back that up (thanks to early adopters of its CGM-based ‘Cyborg’ wearable).

“Different platforms optimize for different things. Oura, for example, optimize for sleep. Whoop provides for recovery. And here we’re optimizing for metabolism,” argues Kumar, adding that how Ultrahuman captures data (with “more real-time” sensors) is a distinctive technical element of its differentiation vs rival smart ring makers.

“The way we have built the data pointers, the frequency of data pointers, the type of metrics, real-time-ness of temperature etc, is more optimized towards the metabolism than other wearables,” he also suggests.

“For us temperature is a much more important biomarker given that we’re looking at the metabolic rate and glucose metabolism. So that was one of the reasons why we decided to build our own wearable — so that we have control over the accuracy of the insights and also the ability to derive some of these insights which was not possible with the existing class of wearables.”

According to Kumar, the Ultrahuman Ring measures stress by looking at factors like heart rate, HRV (heart rate variability) and temperature — running its own algorithmic analysis of the data to identify a per user stress response.

For activity, he says it’s aiming to identify “activity density” — by looking at input from accelerometers, as well as temperature and heart rate — to try to understand “what zone of activity were you in”.

The sleep tracking component also pulls data from activity sensors, temperature and heart rate — to identify different phases of sleep (REM, deep sleep etc). 

While sensor-laden, the Ultrahuman Ring is not currently configured to deliver direct feedback at the hardware level (such as by vibrations) — but Kumar suggests that haptic nudges and/or smart alarms are something it wants to add in future. 

Two wearables for gut insights

Ultrahuman settled on a smart ring as its choice of form factor for this second wearable, rather than — say — a smart band, for a few reasons. Firstly, it avoids the risk of having to compete with existing wrist-mounted wearables (like the Apple Watch) for space on the user’s person. But Kumar also says its testing showed that a ring form factor yielded the lowest data variability of all the forms it tested for metrics like temperature, an important consideration for accuracy.

The team also judged that a ring stands a better chance of being worn more consistently and continuously than other types of wearables. (Ultrahuman’s Ring can survive getting wet in the shower or the pool, he confirms, with also up to five days of battery life before it needs to be charged.) “The more data that the user has about themselves, the more powerful the insights will be,” he adds.

If the ring user is simultaneously wearing Ultrahuman’s CGM too, insights picked up by the ring’s sensors will be directly linked to their real-time glucose levels (which the Cyborg sensor measures via changes to the interstitial fluid under the skin of their arm) — enabling actionable connections to be made between glucose variability and lifestyle events which may be triggers (high stress, poor sleep, low activity levels etc).

“Where our strength will be is in terms of marrying things like glucose variability and impact on your sleep,” he predicts. “Or, for example, if you have a late meal and a late glucose spike — what impact did it cause on your sleep?

“For some people this is perfectly fine — they can have a late glucose spike and they’ll actually pretty much be in the [target sleep] zone. But for a lot of people it actually affects their REM sleep pretty [badly]. And in some cases it affects their deep sleep also.”

The Ultrahuman smart ring plus Cyborg CGM combo could therefore power diet-related interventions for users who can’t avoid having a late meal but for whom its metabolic tracking has implicated glucose spikes as negatively affecting their sleep quality — by suggesting, for example, they opt for certain foods that are linked to improved sleep (such as tryptophan rich foods) when they have to have a late meal.

“Those are the insights where we will actually be pretty unique,” he suggests.

The product will also approach movement and activity recommendations in a distinct way to rival products, per Kumar.

“Movement is not just about burning more calories — it’s also about frontal lobe development, it’s also about longevity. And movement is an activity which helps people reduce cortisol levels and at the same time increase their [high calorie] expenditure. So a lot of our focus is going to be around movement — if you look at it from an activity tracking perspective.”

A user of Ultrahuman’s smart ring who has not tapped its upper-arm-mounted CGM sensor yet can still get some general benefits, according to Kumar. But he emphasizes that the greatest utility comes from the combination of the two wearables. “People will be able to understand their sleep quality, people will be able to understand their levels of stress recovery, movement etc. But if they want to understand the effect on their glucose metabolism of all these factors they have to unlock it by a CGM. So it works both ways,” he says.

The ring can also work to bridge service gaps inevitably affecting the Cyborg sensor — and thereby expand the utility of its CGM tracking service — by continuing to provide a prior Cyborg sensor user with personalized feedback after their sensor has expired. (The arm-mounted CGMs typically last two weeks before they have to be replaced — meaning the Cyborg service is interrupting unless a fresh sensor is applied — whereas the Ultrahuman Ring is designed to stick around for longer and won’t automatically ‘expire’ in the same way.)

“[If you just have the ring] the platform will understand, based on what sort of metabolic rate, your carb processing capabilities, how much you should be walking, for example, after a meal,” explains Kumar. “And that’s possible because now we understand what sort of activity levels led towards decreasing throughputs. So that’s how — over time — we actually don’t need your CGM data also, in many scenarios, to derive this output.”

Bringing biohacking to Boomers?

Since the ring form factor is obviously more accessible vs the (semi-invasive) arm-mounted CGM, Ultrahuman is expecting greater adoption of the smart ring than for the Cyborg tracker.

He says it currently has 25,000 people in India on the wait-list for the Cyborg service — which remains in a managed beta — but it’s expecting at least 100,000 people to buy into the smart ring over the next year.

Ultrahuman will be selling the smart ring globally — whereas availability of the Cyborg sensor remains limited to India and the UAE, owing to regulatory considerations and also its decision to focus on markets with high rates of metabolic disorders for the product to target — so the pool of potential buyers is larger.

At the same time, Kumar says the team is hoping the smart ring will be able to act as a broader marketing tool to cross-sell the CGM-powered service.

The typical profile of existing Cyborg users is an individual between 30-40 years old with a passion for fitness (and/or data analysis), and an interest in preventative health. But with the smart ring expected to have broader appeal, Ultrahuman now has its eye on convincing older, Baby Boomer generation consumers to take a punt on its metabolic health service — a wider population (of circa 25M-30M globally) that Kumar suggests hasn’t adopted a health wearable as yet. But maybe a shiny bit ‘o’ bling could be just the nudge they need…

“Maybe they have adopted a wearable like Apple Watch because it’s not just a wearable for health — it also does a bunch of things — but they haven’t gone deep into deep health or a biohacking wearable yet. So that’s what our target audience in the future would be — but the first audience is going to be biohackers, people who love data about their health,” he adds.

 

US medical device maker, Abbott, is moving into making general purpose consumer biosensing wearables.

The company has been making continuous glucose monitor (CGM) hardware for diabetes management for years (since 2014) — but in a healthtech keynote at CES yesterday, Abbott’s chairman and CEO, Robert B Ford, announced it’s developing a new line of consumer biowearables — called Lingo — intended for more general fitness and wellness purposes.

“Technology gives us the power to digitize, decentralize and democratize healthcare, create a shared language between you and your doctor — and put more control of your health in your hands,” he said during the keynote. “We’re creating a future that will bring you and your loved ones care that’s more personal and precise. It’s happening right now. And its potential is no less than incredible.”

Ford said the Lingo sensing technology is being designed to track “key signals” in the body — such as glucose, ketones and lactate — adding that it could also be used to track alcohol levels in the future.

Last year the company launched a biosensor designed for athletes, called the Libre Sense Glucose Sport Biowearableiii — which was made available in Europe, and has been used by the likes of marathon world record holder Eliud Kipchoge to support their training needs.

Abbott said its goal with Lingo is to expand glucose monitoring to people looking to manage their weight, sleep better, improve energy and think clearer.

To support this expanded utility it said it’s developing the biosensor to measure other biomarkers than glucose.

“A ketone biowearable is being developed to track ketones continuously, see how fast you are getting into ketosis, and understand exactly what keeps you there by providing insights on dieting and weight loss,” the company noted in a press release. “A lactate biowearable is in development to track continuous lactate build up during exercise, which can be used as an indicator of athletic performance.”

In recent years a number of startups in the US, Europe and Asia have been seeking to productize CGM hardware — including existing sensors made by Abbott — for a variety of non-medical purposes, launching real-time blood glucose tracking services targeted at fitness enthusiasts, peopler wanting to lose weight or generally health conscious consumers.

Abbott jumping into the space itself so quickly suggests it sees significant potential for biosensing consumer wearables to go mainstream.

For a deep dive on what it’s like living with a CGM biosensor attached to your arm — and the constantly updating window into biological process that it provides — check out TechCrunch’s review of Ultrahuman’s Cyborg service, an Indian-based startup that’s repurposing current-gen sensing hardware made by Abbott.

 

For four weeks during 2021, this TechCrunch reporter took the plunge and tested a “metabolic fitness” service from Bangalore-based startup Ultrahuman. The tracker program, branded Cyborg, uses arm-mounted medical grade hardware to get a real-time read-out of your blood glucose — using that dynamic data-point to power a quantified health service that scores what you eat and how you move, nudging you to make healthier lifestyle choices throughout the day.

Research has linked chronic metabolic inflammation, from factors such as poor diet and physical inactivity, to the risk of developing a number of diseases — from diabetes to cardiovascular disease, chronic kidney disease and even cancer. So the theory behind the product is that lots of incremental lifestyle choices can stack up to a healthier long term outlook — if you’re able to ‘optimize’ these decisions to avoid triggers for inflammation and oxidative stress.

Here follows my long read on the curious experience of living with a skin-perforating wearable and a dynamically updating digital window onto your biological process, as well as wider discussion of the value of continuous glucose monitoring (CGM) for a general health/fitness purpose, and — finally — some notes on the competitive landscape springing up around productizing this type of sensing hardware.

As this is loosely a review of Ultrahuman’s (still private beta) product/service, I’ve also included a ‘Verdict & Price’ section too. Skip ahead if you want to dive into the operational details. But first some context…

 

Preamble & Caveats

Becoming a cyborg is no longer as sci-fi as that sounds. For years the ‘quantified self’ trend has spawned all sorts of sensors and services for measuring bodily activity and nudging you to track and ‘optimize’ your outputs — from step counters and heart rate monitors, to stress and sleep sensors, lung capacity scorers, and, more recently, freakier stuff: Blood glucose monitors and saliva or pee/poop analyzers — the latter for delving into hormonal and/or microbiome/metabolic health if you’re so inclined.

Serving the worried well with wrist-mounted, strapped on or otherwise self-administered sensing technology plus a subscription service to play pocket oracle — via an app-delivered interpretation of what all this personal data means (and ofc how to improve your metrics) — is booming business. Close your (exercise) rings. Breathe more deeply. Try to get to bed earlier, and so on.

Some of this quantified health tech can come across as a bit superficial or frivolous; an attempt to ‘gizmoify’ daily life and push a gadget when you could just go for a walk or get to bed earlier. The more basic products work by selling the motivation-challenged a call to get off the sofa or a replacement for lost childhood structure. Or, well, data-fication as proof of existence.

But it can be horses for courses, too; if you have a sleep disorder or suffer from stress and anxiety then tracking your sleep — and getting little nudges and tips on how get more shut-eye — might be just what you need to lock in quality Zzzs.

Available tech has been getting more sophisticated, too. Although, when commercial trackers put a suggestive focus on organ-function (heart; lung etc), the quantification may sound impressive but can suffer from questionable accuracy — given a lot of this stuff is consumer-grade, rather than (regulated) medical devices.

Even step tracker data can be plenty inexact.

But in a more recent development, a growing number of startups are making use of medical grade sensing hardware to offer self-administered metabolic analysis via tracking (near) real-time changes in blood glucose through the use of a sensor that you ‘wear’ on (and, well, in) the skin.

This is a fascinating and growing but still novel area of focus for quantified health startups. One that looks promising, in terms of being able to serve individually useful health insights and which — given enough data — may be able to scale in utility and help empower many others to make healthier individual lifestyle choices.

But the really big caveat is that scientific understanding of metabolic fitness isn’t yet as complete and holistic as we might hope.

Much is not understood — such as why there can be so much variation between individuals’ metabolic responses (different people eating the exact same diet can have very different responses, for example); or the exact role of inflammation in the risk of developing diseases like diabetes or cancer.

So the ability of startups to play oracle here is bounded by the need for more research. (Albeit, grabbing data to advance research and understanding is a key part of the opportunity entrepreneurs are spying.)

Nor is the sensing hardware in question regulated for the ‘general wellness’ use-case most of these startups are pursuing.

Which means these services remain novel — aka, experimental — even if the hardware they’re repurposing is legit, in the sense of being manufactured by established medical devices firms, and regulated for narrower use (i.e. diabetes management).

Typically these sensors have regulatory clearance for people with diabetes to track their blood glucose — instead of having to do constant finger prick tests. That may lend credibility to startups hooking into the same device makers’ APIs to grab the same data stream. But the interpretative spin such services put on the data is just that: A spin.

Any wider analysis — including lifestyle recommendations — are definitely not FDA approved.

The debates that have continued to rage back and forth for years around nutrition — all the fad diets, bestselling books and rehashed discussions of what’s good or bad for us to eat, or even what’s effective exercise — is a long-running symptom of a still flawed understanding of the interplay between our biology and what we routinely expose it to.

It’s clear that measuring complex systems without a full understanding of how all the constituent parts can interact and interplay means you’re not going to get the full picture. At best it’s a snapshot — maybe one that supports improved understanding. But it’s never going to have all the answers. So, another word of caution, the risk of misinterpretation is real.

There is also the question of how exactly do you go about measuring ‘metabolic fitness’? As a label it’s a bit of a fuzzy umbrella — arching over complex biological interactions linked to chemical reactions which generate energy in our bodies that may (or may not) mean we’re easily able to maintain a healthy weight; or which can otherwise support or work against us achieving a high level of physical fitness.

What you eat; how; when; and how active and well rested (vs stressed) you were at the time are just a few of the dynamically varying factors that can affect metabolic function. (One illustrative example: What you ate the day before may affect how your body metabolizes a particular foodstuff today.) While the biomarker (or biomarkers) a product chooses to zero in on and track will also, obviously, influence what that “metabolic fitness” service can see — and is able to deduce.

Startups targeting metabolic health are exploring a range of options — from tracking blood glucose, to analyzing the gut microbiome or other bodily excretions (like urine), or looking at a combination of outputs/signals (maybe also factoring in heart rate). Over time more bodily signals are likely to be added to the mix to try to flesh out a fuller understanding — but a lot of the current gen metabolic tracking is best thought of as a piece of the puzzle; a sketch or a rough guess, with more blanks than shading lines.

How to understand — or, well, best interpret — data from a combination of metabolic signals presents no shortage of questions and challenges for those trying to productize the cutting edge of figuring out all this bodily chemistry. As Ultrahuman’s founder acknowledges — telling TechCrunch: “Solving for accuracy of insights that we generate from glucose biomarkers is at the very core of our mission.”

The company’s website also contains a text disclaimer that the Cyborg service provides “general information for athletes to understand their glucose levels and athletic performance”; and does not substitute for a professional medical opinion or consist of healthcare/treatment for specific conditions or medical concerns.

While an entrepreneurial mission to demystify the metabolism — and commercialize the concept of metabolic fitness — remains very much ongoing, a couple of things are clear: 1) Demand to better understand biological function exists (plenty of people, not just elite athletes, are interested in what’s going on with their bodies generally and their metabolism specifically) — and: 2) big but as yet unverifiable claims are being made for what this type of ‘health’ tracking tech could provide an individual user as a long term benefit.

So — another caveat! — anyone keen to get involved with metabolic biohacking needs to be clear about the limitations.

Getting a bit of data is not the same as getting a diagnosis — or even a proper understanding. More data in this context can mean more noise and confusion, not necessarily a clear signal. It may also make you worried about things you shouldn’t.

Another observation: The consumer boom in digital health/wellness tracking over the past decade has been understandably slower on the update when it comes to invasive/semi-invasive wearables. Aka, sensing devices that work by being installed (at least a little bit) inside the body.

Even partially — dipping under the skin so it can stick a sensing filament into the interstitial fluid in the case of Ultrahuman’s Cyborg — the ‘wearable’ metabolic tracking service that’s the main focus of this review. This semi-invasive-sensor-plus-app combo monitors (near) real-time glucose levels as a proxy for understanding and scoring metabolic health — providing the patch-wearer with blood sugar-triggered nudges and alerts to encourage beneficial lifestyle tweaks.

The goal is to support the sensor-wearer to stabilize their glucose levels as they go about their day — avoiding extreme highs or lows — with the overarching mission of reducing inflammation and oxidative stress, which is linked to negative health outcomes.

The suggestion is that, by paying attention to “metabolic fitness” — Ultrahuman’s phrase of choice to describe its mission — and taking little actions related to what you eat and when, and how and when you exercise and sleep — you can avoid or even reverse chronic inflammation that might, over time, lead to developing a metabolic disorder like diabetes, or non-alcoholic fatty liver disease or cardiovascular disease.

While the notion of dieting isn’t always overtly promoted by startups productizing CGM technology, blood sugar spikes are also of course associated with the consumption of sugary foods (and with a higher volume of consumption) — both of which can lead to weight gain. So supporting metabolic fitness implies help to obtain and stay a healthy weight too.

With such headline-grabbing potential gains — from reducing the risk of chronic diseases to support for weight management and a smart digital sidekick to boost athletic performance — it’s easy to see why there’s been a major startup scramble to demystify (and monetize) the metabolism.

And when it comes to startup opportunity, a literally ‘wired in’ consumer health tracker is definitely a lot less mainstream than wrist-mounted tracking gear like the Apple Watch — which instantly shrinks competition from consumer tech giants, providing bold entrepreneurs with a chance to shine.

Safe to say, if Apple’s wearable came with a retractable metallic fang embedded in the backplate, the tech giant wouldn’t have shipped anywhere near as many watches, no matter how fancy-looking its glucose-sensing filament. (Apple will surely want to incorporate a needle-free version of glucose monitoring into its Watch, as rumors have suggested, and if the tech works out.)

Piercing the skin just sounds messy (even if it isn’t really). And ofc lots of people hate the idea of needles. That means there is simply more space and opportunity — here at biohacking’s cutting edge (ha!) — for quantified health startups to blaze a trail that seems to go deeper than mainstream consumer tech companies. Quantified self tech that’s not afraid to cross the needle-phobia line certainly feels more serious because it is literally closer to the biological process that’s being tracked.

That said, whether placing a tracker into the skin makes a meaningful difference vs a less intimate sensor placement — in terms of the quality of the data being captured; the analysis of that data; and any resulting recommendations provided to the user — is a tricky-to-answer question. (Indeed, it’s a whole series of questions, depending on context and the execution of the service/s.)

In the case of Ultrahuman’s Cyborg, the startup is careful not to overpromise; its marketing puts the responsibility on users to “work on improving your health with real-time visibility of how food and exercise impact your body, and a score that motivates you to improve every day”, as the minimalist instructions which arrived in the box with the beta product put it.

The metabolic “score” Cyborg gives is personalized, yes, But it’s an abstraction and interpretation of biological processes that still hold plenty of questions for science. So, again, this is really more about being part of a search for answers vs getting a single ‘biological truth’ handed to you on a plate. (In short, there is no plate; there’s just a lot of suggestive data to feast your curiosity on.)

While a little knowledge can be a dangerous thing — maybe even more so when the data in question is attached to your own biology — a glimpse of one’s inner workings is undoubtedly catnip for the curious. And in our digital age, with so much health research information available online at the stroke of a key, who isn’t a little curious on matters of personal biology?

The danger, perhaps, is that a more invasive/intimate sensor placement may lead people to automatically assume this type of tracker is giving higher quality intel (and more personally relevant insights) than the calibre of the data processing — and our wider scientific understanding of metabolic processes — is able to deliver.

Ultrahuman isn’t afraid to push that sci-fi notion as a selling point, though. Hence its overt choice of ‘Cyborg’ branding — which deliberately emphasizes the intradermal sensor placement — the direct interface between the tech and your body — implying that’s the special sauce powering a quantified health service which promises to “nudge you towards better health, one small step at a time”, and without the need for “drastic diet changes” or the tedium/frustration of “generic exercise plans”.

Given how many other startups are also leveraging the same (or similar) CGM hardware, the automagic of obtaining the data is already at risk of being commoditized; it’s how this information gets visualized, analyzed and contextualized for each user that really counts.

Again, though, given all the aforementioned uncertainty around the science, that looks inherently hard to quantify.

Of course cynics might say that makes for a perfect startup opportunity…

---

Ultrahuman Cyborg: How it works

Tracking dynamic changes in blood glucose is Ultrahuman’s proxy of choice for assessing metabolic health.

Why glucose? Ultrahuman’s CEO and co-founder, Mohit Kumar, says it was the best fit for what they wanted the product to achieve — being “a real-time biomarker that is sensitive to food, stress, sleep and activity”.

“We were looking for biomarkers and methods to personalize the fitness journey for people when we started as well but it took us a year long of experimentation to figure out which biomarker really works for the kind of impact we were looking at. We looked at various biomarkers like HRV [heart rate variability], sleep and respiratory rate but glucose seemed really interesting out of the entire lot because of the feedback it provides on our food aspect of the lifestyle,” he tells TechCrunch.

“This means that we would be able to get instant feedback for these lifestyle factors and what we’ve seen is that instant feedback leads to better actionability. For e.g. a nudge that pushes you to walk after a meal that gives you a spike will lead to better actionability vs a report that gets sent after a day.

“Secondly, there are so many fitness wearables and markers that help you improve your activity performance but there’s nothing that helps you optimize what you eat. Nutrition is generally a blackbox and is way more confusing (given hundreds of diet types and personal preference) but it is probably the most important lifestyle factor given how broken our food ecosystem is.

“This is why we felt going in with glucose makes a ton of sense from a ROI perspective even though it is a semi-invasive biomarker. The private beta is helping us understand what nudges and information helps people make lifestyle changes easily. We’ve seen massive engagement on the platform with app opens / user being around 21 per day and most people seeing real-improvements in their health around the 45th day of usage.”

Tracking blood sugar swings almost as they’re happening — thanks to CGM technology — is immediately a major step up on the patience-challenging business of (traditional) dieting trial and error over a multi-week/month period: Aka, change what you eat/how you exercise and wait and see if it actually moves the scales, weeks or even months later.

Continuous blood glucose tracking (vs repeated finger pricks) has been enabled by the development of CGM hardware in recent years — initially for people with a formal diagnosis of diabetes. But now a growing number of startups are productizing this technology for a more general health-concerned or fitness-focused consumer.

The tech itself has led to some interesting science. See for example this 2018 research paper which showed that glucose dysregulation (i.e. highs or lows outside what’s considered the normal range) were actually pretty common in ‘healthy’ people; i.e. those without a diagnosis of diabetes or pre-diabetes — which wasn’t what the researchers had been expecting to find.

At a basic level, Ultrahuman’s service consists of arm-mounted sensing hardware (a disc-shaped sensor) — which must be replaced every two weeks — plus an app to visualize your blood glucose data and deliver alerts and nudges. You pair each new sensor with the app to continue the tracking.

The sensor hardware is made by another company: US medical devices firm, Abbott. The specific sensor that shipped with the Ultrahuman product at the time of writing was Abbott’s FreeStyle Libre 2 flash glucose monitoring system.

Self administering the CGM sensor is a little nerve wracking but mainly because you only get one shot at (correct) placement. And with only two sensors in the beta box delivered to TechCrunch I didn’t want to waste any hardware.

At the time of application, Ultrahuman had produced a couple of (amusingly robot-voiced) videos to instruct on sensor placement and set up. These were helpful — and only slightly disturbing (owing to the reference not to press too hard to avoid “few drops of blood splatter”).

Abbott’s hardware comes with its own set of instructions and a spring-loaded applicator which you prime manually before positioning the plastic cup on your raised upper arm and, trepidatiously, pressing down to fire the filament into your flesh. The action is rapid enough to make you flinch. It may not help to recall another phrase in Ultrahuman’s instruction video (“hollow needle”). But the needle is just the delivery mechanism for the filament; you’re not going to be left with that bit of visible metal in your arm.

Was there any blood splatter? Not that I noticed. However the second sensor I installed/put on seemed to have fired into a nerve or something as it was pretty painful for several days. After which it sort of settled down/bedded in. Or, well, I got used to it.

The first sensor was not painful, per se, to wear but it definitely takes a bit of getting used to to sleep with a piece of plastic attached to your arm. I found certain yoga poses required extra contortions to avoid uncomfortably pressing down on the sensor, for example. And I swear I could hear a very high pitched whine in my head at night while wearing the CGM — but maybe I was just dreaming of electric sheep.

Yes you can shower/bathe with the sensor in place. Ultrahuman’s box contained a few disc-shaped fabric tape patches to help protect the hardware (and add its branding to your arm). These can start peeling off after a few days depending on your lifestyle but the sensor itself remained firmly lodged for both my two-week stints. (You can remove a dogeared patch and replace it with a fresh one (if you have enough spare). Although that was also nerve wracking as you don’t want early patch removal to prematurely rip out the sensor. So basically it’s about as much fun as applying a whole Macbook decal.)

If you’re curious about the sensing filament itself it feels like a piece of not that fine wire. You get to see it for the first time on extraction from your arm. At which point I saw it looked as if it was coated in some kind of black paint. Which was — I was not too pleased to observe — flaking slightly… But by that time you’ve been living with it in your skin for two weeks so Cyborg acceptance has already taken place. Smart.

Was there a mark left? Yeah, a small red bump where the filament had perforated the skin. It faded after a while. The tape itself — including the sensor’s built in fixing (which stayed a lot more firmly attached) — never bothered me.

The sensor pairs with Ultrahuman’s app via Bluetooth. This means it can lose connection if your phone is out of a few meters’ proximity with your arm/person — at which point the data flow (and real-time alerts) will stop. So now you have the perfect excuse for your phone never to leave your side!

If that does happen, the app will notify you and request you to tap the phone back on the sensor when you can to upload any missing readings. (NB: On set up, the sensor also needs a little time to “warm up” — before data starts flowing. So you may find yourself pacing the room as you wait for it to be ready to log your first workout/meal etc.)

The app itself was a work in progress during TechCrunch’s period of testing which was split over more than a month (as I took a break between applying sensor 1 and sensor 2) — and the software went through a number of changes, including one major visual tweak.

This changed the glucose plot line’s gradient from a too simplistic view (where low-to-high glucose was always displayed as green-to-red) to having a central “target zone” where the plot displays in frosty ‘good to go’ green but as/if rates drop too low or too high they will bleed in a gradient from yellow to orange to red — meaning you can have red highs and lows if you’re out of the optimal glucose range (which is between 70mg/dL and 110mg/dL).

This update was a vast improvement as the earlier version had been visually suggesting that a lower glucose was always better — even if the level was already below target (aka, hypoglycemia) — which it just a small illustration of some of the design/UX pitfalls for this type of quantified health product.

As well as plotting the ups and downs of your blood glucose throughout the day (or at least the approximation which Abbott’s hardware pulls from your interstitial fluid; as any diabetic could tell you, these levels don’t always exactly match blood glucose readings; and if your glucose is rising or falling there can be a short time lag before that shows up in a flash glucose monitor), the app displays what Ultrahuman refers to as a “metabolic score” — which is a number from ‘0’ to ‘100’.

This is the main ‘metric’ mechanic the app uses to try to nudge and gamify healthy lifestyle tweaks.

Ultrahuman describes this score as an indicator of your “overall metabolic health” and says it’s calculated based on glucose variability, average glucose and time in target metrics. The number resets to 100 every day at midnight and decreases or increases “based on your daily lifestyle activities and body’s response”.

The gamification mission sounds very simple: “Your goal is to maximize this score on a daily basis,” as the app puts it.

In practice, getting a ‘good’ (i.e. high) score will depend on your individual biology and lifestyle. And, depressingly, you can wake up with a score that’s already down in the 80s (or, I guess, worse) — depending on what you did/ate earlier.

NB: Stress can also impact blood sugar so events outside your control can impinge on your metrics.

Logging of food — and/or activity or the other types of events which were gradually added to the app during the testing period — is done manually.

Initially this was by custom typing your meal descriptions (or activity). A later update added a food and activity index that lets you search and pick from a structured list and their quantities or times rather than manually typing everything.

In the end, I much preferred custom typing to log food as the list was far too specific and tedious to feel useful. (Type: “Cheese” and it’ll suggest a lot of different types of cheeses — but not necessarily the exact one you’re eating, nor the amount you actually have on your plate, which you may not know in any case; and that’s just one meal ingredient to log; repeating that for a full plate quickly gets old fast… Plus the list also seemed pretty US-centric, which wasn’t very useful for logging a European diet.)

Whereas type in your own favorite cheese — or indeed a custom description of the entire meal — and you can quickly log it next time you eat it as the app will remember your custom label.

Doubtless Ultrahuman is keen to get the best quality structured data that it can — to build the wider utility sought for from predictive AI models. But, if logging feels too much like work, few users will perform the task perfectly for free. So it may have its work cut out to get accurate and structured (vs custom but cryptic) food-to-glucose-response data from its beta user-base.

(Indeed, it may need to rely on asking users to snap an image of their meal and applying computer vision technology to make informed deductions, say, although that may also introduce plenty of errors. Longer term — if the tech goes really mainstream — you could imagine restaurants printing a QR code per meal on their menus which can be scanned so all the correct nutrient data is instantly logged to reduce input friction.)

Activity logging was a lot more straightforward than meals. Not least because, unless you’re an Olympic athlete, you’re going to need to log a lot less of it than food.

After community feedback from beta users, Ultrahuman also added “stress” events as a logging option, as well as fasting — which can of course play havoc with blood glucose but — per some research — may have its own set of health benefits. So giving users more granular options to help better structure the CGM data is sensible.

In the future, automated logging via integration with other types of consumer wearables seems likely. For example, it’s easy to imagine that your fitness band or smart watch detects a specific activity and passes that data to the Ultrahuman app — which could just prompt the user to confirm the details of the detected workout.

For now, though, beta users remain in control of inputting and structuring the data so data quality is likely to be a real smorgasbord.

---

 

What I learnt

Ultrahuman’s tips caution that during early use of the product you probably won’t be rewarded with stable, high scores.

This is because learning what you need to do to stabilize your blood glucose typically takes a bit of time — since you have to try different stuff (food pairings, exercise timings etc) to see what works for you. Although this is still a much accelerated process vs the tedious business of old school dieting and fitness regime assessment. (Ofc if you’re blessed with a naturally more stable (i.e. low variation) glucotype you may find you need to do a lot less manual ‘steering’, as it were.)

I will still unprepared for the early horror, though. And spent pretty much the whole first week — jaw on the floor — watching the app lowball score the stuff I usually eat.

Humus salad pita bread sandwich lunch followed by a few walnuts, half an apple and coffee (white, no sugar)? Doesn’t that sound reasonably healthy? Um, apparently not, in my case. It remains one of the all time “bottom zone” lunches during my four weeks as a Cyborg (scoring a big fat ‘0’).

The all time worst lunch food during the test period (purely in terms of how high my glucose peaked after eating) was at least not a dish I had prepared myself but a fast food meal — albeit, from a brand that markets its fare as a “natural” and more healthy choice than traditional McBurgers n chips.

The meal in question — Leon’s lentil masala with brown rice followed by a “regular” coconut milk latte (brand of plant milk: Rude Health); another ‘0’ scorer — produced such an epically large spike that I decided I needed to do an emergency HIIT session just to bring my elevated levels back down again.

The exercise did do the trick. However, if I’d known before lunch that I would need to do do a bunch of burpees and squats right after lunch to metabolise the food spike I might well have revised my food choices.

How continuous glucose monitoring, and mass access to real-time metabolic data, will affect the fast food industry is certainly an interesting question to ponder…

I hadn’t checked the ingredients small print prior to eating the Leon meal — but eyeing the label suspiciously afterwards, in the red glow of the app’s condemnation, I was unimpressed to find “caster sugar” in a long list of additions.

Although, knowing what I know now, it was probably the coconut milk (an ingredient in both the stew and the coffee) that was especially triggering for me.

Sadly, the post-meal coffee probably didn’t help either.

My least favorite Cyborg learning was that coffee seems to raise my blood sugar. Green tea? Totally fine. Black coffee, decaf, white? All cause me some uplift. And since I like to drink coffee in the afternoon, after eating lunch, that means a raise atop a (food) raise — which might be just enough to tip me into the red.

I still refuse to be a morning coffee person, though.

Rice can also be a spiker for many people — certainly white rice which is more quickly metabolized by the body vs the more fiber-rich wholegrain. Although I’m now more wary of the crash that can come after eating a mainly white rice-based evening meal as it seems to work against keeping blood sugar stable and sustained in the target zone overnight.

Blood sugar lows are just as important to avoid as highs, as it turns out. At least, that’s my sense after four weeks hooked to a CGM. Although my early usage of the app was entirely preoccupied with trying to avoid the big red spikes, they did get easier to manage over time — with some creative biohacks and a few strategic dietary edits.

For example, I have all but removed plant-based milks from my diet (save for a dash of oat milk in coffee; no I have not — will not! — give up coffee entirely. But I do tend to nurse a cup for longer now). The spikes these alt milks served up were just too consistently red flag-ish to ignore and I came to think of them as akin to fruit juice and best avoided. Which — again — is pretty interesting considering how often the marketing of these highly processed beverages makes loud noises about how they offer a ‘healthy choice’.

Interestingly, other Cyborg users seem to have reported a similar issue — per one of the company’s email newsletter round-ups of shared learnings, where it wrote that: “Almond milk and breakfast cereal could actually cause a bigger spike than a hotel breakfast buffet!”

Maybe this is a similar mechanism as can cause a glass of orange juice to spike whereas eating a whole orange (typically) won’t. Or maybe it’s down to something more specific in how these drinks are manufactured — the type of processing they undergo and specific additions. Many have added sugar for instance (although the ones I was pouring on my cereal definitely didn’t — yet they still spiked me). Unfortunately I didn’t have a chance to make a homemade version of oat milk to do a direct comparison with commercial brands to see if it was any less spikey.

For breakfast I do still usually eat a bowl of oats — which certainly also has spike potential (being carbs, albeit fiber-rich carbs) — but I make sure they are jumbo oats (not oatmeal). Most importantly, I liberally dust the bowl with cinnamon (which I discovered helps reduce glucose spikes). And I eat them with water (not any kind of milk), plus a blob of natural yogurt (for flavor and some essential vitamins), plus the usual mix of berries and seeds.

This is not a massive change on my pre-CGM breakfast of choice (oats, berries, seeds etc but washed down with, er, oat milk). But the difference in metabolic score terms? Huge! It switches from a meal that typically scores a ‘2’ to a ‘9’. Crazy but true (or, well, true per Ultrahuman’s reading of my fluctuating interstitial fluids). 

I also found creative ways to adapt how I consume bread to limit how much of a spike it generates.

Eating less or even no bread is one way to shrink glycemic load and manage down any associated blood sugar rise. However like oats, wholegrain bread is a complex carb that has dietary benefits so I didn’t want to remove it (or, indeed, quit carbs entirely) from my diet. So, with the benefit of the app’s real-time glucose view, I experimented with eating a slice of wholemeal bread towards the end of lunch, after other fiber, protein and fat rich foodstuffs — which take the body longer to break down — and that seemed helpful.

I then found another specific biohack — involving apple cider vinegar — that worked a treat.

As with cinnamon, I learned this type of fermented vinegar has properties that help to reduce glucose spikes. So I experimented with pouring the vinegar (stay with me) on a slice of sourdough bread before eating it — yes this sounds odd but actually tastes amazing! Using this method, plus eating the bread later on in the meal (after the salad, nuts etc), I could turn a lunch that spiked into one that remained in the healthy zone. There’s simply no way I’d have figured out something as specific as that without being able to see real-time shifts in my blood sugar.

The problem is the lunches that spiked didn’t make me feel any different/less healthy vs the lunches that didn’t. Not without seeing the metabolic response in the app. So it just wouldn’t have been possible to distinguish between them without the sensor data.

Plus, of course, another person, with different metabolic responses, may be able to eat five slices of bread without any spikes at all. So there really is no clever way to generalize — beyond setting basic strictures such as control your carb intake and carefully build the balance of foodstuffs on your plate. And generic, broad-brush strategies that can be very demotivating in the absence of immediate feedback — which is exactly what makes the CGM so potentially, individually transformative as a lifestyle tool. Suddenly you can try stuff and see if it actually works for you or not.

That said, whether managing relatively small blood sugar spikes is as important for a person’s long term health as metabolic tracker startups like to suggest is a wider question.

Dr Matthew Campbell, a scientist who does research into biological systems that impact the human metabolism at the University of Sunderland in the UK, was sceptical about the benefits of otherwise ‘healthy’ people putting so much effort into managing their blood glucose when we asked for his views on this general use of CGM technology.

“Glucose usually fluctuates throughout the day anyway — it’s not a kind of static variable, it is very dynamic. But it should, on average, stay within a normal range. There are cut off points for people who would be characterised as high risk. For example, if your glucose doesn’t come down below a certain level after a meal or in the morning time if it’s chronically elevated. And that’s where the kind of cut points are for diagnosing diabetes or even pre-diabetes, the people who are at risk of developing diabetes,” he tells TechCrunch.

“The issue that we have [with ‘healthy’ people tracking their glucose] is just these arbitrary values — if it’s going down that’s okay, if it’s going up that’s not so good — [but] if you sit within the normal range I don’t know what the clinical utility and the usefulness or the health advantage is of, for example, reducing your glucose by 1mmol if it’s already in the healthy range.

“So I guess if you already sit — 95% of the time — within a healthy range trying to flatten that line or aggressively manage it even lower, I don’t think that confers any additional health benefit because you are already in a healthy range.”

Campbell also pointed to the challenge of correctly linking the blood glucose data that comes from the CGM to everything going on in the user’s body which might be influencing glucose levels, noting too that as well as a time lag the exact position of the sensor on the user’s arm can affect the readings, for example.

“So certain situations, is it your weight, your sex, your ethnicity, individual genetic makeup — all of those different factors influence glucose levels — sleep impacts it, nutrition impacts it,” he says. “And I think if this tech just [tracks] the glucose trace and it doesn’t tie in those other factors then it’s quite difficult to make an informed decision on what is influencing your glucose levels.”

He was more positive about the potential of CGM for athletes, though.

“I think what it can be useful for — you mentioned elite athletes — if you’re exercising at particularly high levels or for a long duration of time — even if you don’t have diabetes, can be at risk of having low blood sugar levels and a lot of this tech tends to come with alerts,” he adds.

Campbell also raised an interesting comparison — suggesting out of range glucose may not always be a problem if the individual’s metabolism is able to aggressively manage it back down again.

“The way to think of it is a little bit like heart rate during exercise,” he says. “If you’re exercising somebody might have a much higher heart rate at the same exercise intensity as somebody else and you might think they’re exercising a lot harder therefore they might be less fit.

“But actually if the variability within the heart rate is a lot greater then that’s more indicative of more cardiovascular flexibility. Which is pretty much associated with very good exercise tolerance and very good levels of fitness — and I don’t really see how it’s any different with glucose response.

“So it’s not necessarily the fact that the glucose level goes outside of range because that happens for a large proportion of people and they can be metabolically healthy — I think what is important is looking at the overall picture.”

Given that, Campbell suggested the true utility of these services will be in augmenting the CGM data with algorithms and machine learning — that can “look for patterns in the data” and “piece things together rather than just cherry picking ‘well your glucose level went high after you did this’; well it doesn’t really matter if it came down fairly aggressively, maybe that’s actually a good thing.”

Returning to blood sugar lows, I had an interesting personal experience in that I was able to figure out — through usage of the app (including by chatting to Ultrahuman’s in-app coaches to get their manual analysis of my CGM data) — that a series of glucose lows I had experienced overnight correlated with waking up in the middle of the night in a cold sweat or even with cramps.

I also noticed that such overnight lows often followed a meal that had involved drinking alcohol (which, turns out, plays its own devilish game of interference with normal metabolic processes). So keeping a careful eye on the ratio of food to alcohol — and perhaps eating a protein-rich snack before bed after an evening meal when I had drunk wine with a less nutrient dense dinner (white rice, say) — was another little hack I was able to work in to shrink the risk of going hypo/crampy in the night without having to forgo wine with a meal.

In that case the personal benefit looks tangible: Not having my sleep unpleasantly disturbed.

I was also able to extrapolate this finding to suggest a similar night time snack hack for an elderly relative — who had been suffering chronic night cramps for months. After she’d adapted her regime to include a strategic bedtime snack she soon reported being almost entirely cramp free overnight.

These are of course just a couple of anecdotal examples — but they are illustrative of the potential for individuals to experiment, make connections and join the dots between the unique quirks of their lifestyle and the CGM data.

Dr Michael Snyder, a Stanford professor and co-author of the aforementioned pioneering research paper — who is also a co-founder of a (rival) US startup, called January AI, which sells its own metabolic health tracking service that’s productizing CGM — is, as you’d expect, evangelic about the benefits of the technology to deliver valuable revelations to individual users.

He actually has Type 2 diabetes — and has worn a CGM to help manage the condition for around a decade at this point — so is well placed to comment on the tech’s utility.

Albeit his personal use, for a specific medical condition, is very different to the general fitness/health use Ultrahuman, January AI and other startups in this space are targeting. But he suggests that broader use of CGM technology could help manage or even reverse the risk of people becoming pre-diabetic or diabetic.

“Right away you learn what foods spike you and what doesn’t — and that just differs from one person to the next,” he tells TechCrunch. “You can actually see people who have glucose dysregulation who might not otherwise know it and this is a big deal because 90% of pre-diabetics don’t know it and 70% of those will go on to become diabetic so one could argue it’s really really valuable to get their glucose under control so at least they can push off becoming diabetic hopefully for a number of years.”

“There’s these kind of hidden secrets in your food — at least they’re secrets to you, they’re probably obvious to somebody,” he adds. “But even people who think they knew everything learn stuff, from what I can tell, that they didn’t realize. And, yeah, there’s just sugar everywhere.

“It probably lines up to the concept that I think now — compared to right after World War 2 — people eat something like 40,000x more sugar than they used to. It’s just everywhere.”

[gallery ids="2253113,2253127,2253114,2253112,2253115"]

“I personally think — from my standpoint — the whole world should be getting measured on this at least on some treatments,” he also tells us. “If your glucose is under control maybe you get measured a little bit less, get measured periodically. But if you’re pre-diabetic or diabetic I think this information should be life-saving on some level.”

Snyder also predicts the tech will get a lot more powerful — thanks to the addition of AI and predictive modelling around food responses based on all the empirical data that’s now being ingested after being fed in by early adopters.

“That’s why you need AI,” he notes. “First of all you’ve got to know which foods spike you — which ones don’t. It’s very empirical unless you just do it you don’t know going in — so we’re finding some people spike to grapes, other people to pasta. Everybody spikes to white rice.

“But different people do spike to different things and at some point we’ll get predictive about what’s doing that but right now it’s just empirical. And so that’s what these devices do — they teach you.”

“For January AI we have food recommender system because we can say well here’s what you’re eating that spikes you and we know the composition of these other foods and with reasonable predictive accuracy we can say well this food didn’t spike you, eat that one, don’t eat that one,” he adds.

“It sounds crazy — but it is a big data problem. You need to have a lot of data and a lot of understanding to be able to do that.”

January AI similarly factors in the user’s activity level — given it also impacts glucose level. And Snyder argues that even just tracking those two elements is enough for such a service to be useful.

“I think that’s essentially at least two of the ingredients — but you’re right there are a lot of factors and that’s why it’s a data problem,” he adds. “Bring in enough data around you personally and we’ve got the data to decide what formulas are working for you.”

Personally, I can say one thing for sure: I have never known a gadget to be so engaging. Just on the pure information level.

The Ultrahuman app’s fairly formulaic alerts — which might pop up to warn you that your glucose is rising and suggest you “get movin'” to bring the level down; or nudge you to eat earlier in the evening for “better sleep quality and metabolic response”; or offer some motivation by trumpeting an “epic/insane start to the day” based on minimal spikes/crashes — were probably the least personally useful element of the product for me. Because, well, if you’re paying attention to the data you’ll soon realize that sort of stuff yourself.

I was very quickly way down the rabbit hole of testing diet/exercise tweaks to see whether I could identify hacks and strategies to keep things frosty green.

It’s absolutely fascinating/terrifying to watch how your body deals with the stuff you throw at it. But, be warned: Your S.O. will hate you as you inexorably whip out your phone at lunch/dinner to first log your meal and then vicariously observe as the app scores your body’s response to whatever you’re eating. It’s a double whammy for screen time. And the stickiest app I’ve used since forever. (Sometimes literally given you’re logging what you’re eating.)

But of course it’s not perfect.

One notable functionality issue I found is that the app wasn’t always able to distinguish between an exercise-related spike (yes intense exercise can raise blood sugar out of the target range!) and a food related spike (even if you’re doing careful logging) — so it could end up scoring your day badly when it shouldn’t.

Exercise spikes are “nothing to worry about”, per Ultrahuman’s coaches — who I quizzed about this via the app’s chat function. “The reason for spikes during strength and HIIT workouts is due to an increase in adrenaline and cortisol which stimulates the liver to break down glycogen into glucose,” was the explanation I got from one of the coaches, along with the reassurance that this is: “Nothing to be worried about. It’s natural phenomenon.”

Now a person with diabetes may need to worry about going out of target even if exercise is the cause — as their body could have trouble bringing the elevated blood glucose back down again. But a person without that diagnosis — the more general consumer that Ultrahuman is targeting for Cyborg — shouldn’t, in theory, be worried.

However the app, in its current form, ended up causing me some concern when I did some intense exercise and then right afterwards ate a meal. High glucose rates caused by the HIIT — which the app will normally notify as “a good spike” — seemed to get co-mingled with the food-related increase and that combination conspired to dent my metabolic score.

Accurately distinguishing a “good spike” from a bad one is evidently a work in progress.

Here’s what Kumar told TechCrunch when we asked about this: “Solving for accuracy of insights that we generate from glucose biomarkers is at the very core of our mission. If we look at clinical grade parameters that determine how one’s body responds to something like food, we get to know that it is a combination of: ‘X ( macro+micro constituent of food ) + Y ( the state of recovery i.e stress, sleep deficit, microbiome diversity etc ).’

“The platform currently looks at X closely and hence you would see that there are many exceptions to how glucose responds to food. With our custom hardware that’s going live in early 2022, we are changing the way we look at this by capturing the rest of the Y factors i.e HRV, sleep etc. We feel this will completely change how we look at the food and activity response and the resulting accuracy.

“For e.g.: The platform will be able to clearly figure the attribution of activity and food within a spike. This is because we could figure out your approximate glycogen release thresholds based on a combination of glucose with other factors that we will capture via our custom hardware wearable.”

Kumar also said the startup is starting clinical trials for a study that relates glucose, insulin and other bodily parameters (“Triglycerides and hormone balance”) to establish what he described as “a proper correlation between glucose monitoring predictions (‘metabolic score’) and the actual state of metabolic health”.

“This has also been attempted in the past with lesser tools and non-continuous glucose at disposal but the v2 here will have way more validation,” he predicts.

So, once again, more research is needed to try to improve the resolution of the ‘personalized’ snapshot of data the CGM is pulling out of your arm. Which also means that these cutting edge quantified health services may still be making a relatively crude assessment of what’s going on in your body at any given moment.

There’s a similar complication with food too of course, unless you’re someone who eats a single foodstuff per meal.

Since most of us eat foods in combination (bundling different ingredients), it’s the combo you’re eating that counts — and, indeed, the order in which you eat different ingredients on your plate can affect how you metabolise them. So the same meal eaten in a different way (or at a different time of day) might go down (or up) differently.

Starting with fiber rich foods (salad, vegetables etc), moving through proteins and fats and ending with (any) carbs — a deconstructed humus salad pita lunch, say — would probably have been less of a low scoring lunch for me than wrapping the same food in bread and eating it the quick and convenient way.

Another clear takeaway from my four weeks as a Cyborg is that fast, ‘convenient’ food — scoffed at a pace — will, inexorably, cause big, unhealthy-looking glucose spikes.

I also found that more processed the food (i.e. prepared meals with added sugars, preservatives, oils etc), were more likely to spike vs eating whole foods, freshly prepared.

This was not surprising to me — I’ve long sought to avoid eating heavily processed foods in favor of stuff I prepare myself using fresh/minimally processed ingredients — but it did underscore how much of a problematic food culture the Western world has developed, with its time-is-money emphasis on speed which encourages liberal use of artificial sweeteners and other additives in order to turn an edible convenience food into a profitable product with a long shelf life.

My experience of using a CGM suggests that eating in a way that is healthier for you — because it generates less inflammation and oxidative stress — requires both more time to prepare food and more time to consume food.

Healthier ingredients may also be more expensive to buy and assemble yourself vs buying a product that comes prepackaged and ‘ready to eat’. So health can literally cost more, in time and money. So there are huge socioeconomic considerations when you start to dig into metabolic health.

Cracking open this Pandora’s (lunch)box has implications that scale beyond our broken food system too — touching on wider structural inequalities baked into our societies.

Poor health and poverty are often intertwined. And it remains to be seen whether big data and AI will be able to break that link by democratizing access to valuable health insights — scaling broad utility off of enough individual-level learnings — or whether tech’s wealth divide will just serve to further accelerate inequalities as health tech gets smarter too.

The concept of a cyborg instantly implies a new elite tier of humanity. But what about all those who can’t afford to be wired in?

---

 

Verdict & Price

While I remain (healthily) sceptical of the scale of the potential gains being claimed for metabolic tracking, four weeks as an Ultrahuman Cyborg was long enough to convince me this is the start of something big. And I didn’t have an obvious need going into testing the product — such as wanting to lose weight or needing to get fit. I’m just interested in staying healthy.

Nor am I a big fan of fitness wearables, generally. But this felt like a different level of self quantification.

The future of healthcare will certainly be about shifting towards preventative interventions by leveraging data accessibility to inform and augment our ideas about what’s good and healthy for us — even if, where metabolic health is concerned, there’s no shortage of learning and research still to do.

The data from individual sensors (Ultrahuman’s service alone has some 400x cyborgs at the time of writing) will also feed research that will continue to deepen our understanding of complex metabolic processes. Although there is a degree of risk that commercial interests will look for results which support and underscore their point of view, the potential scale of use — as more of these services launch — should help drive transparency and keep the science clean.

At the same time there is plenty to be cautious about.

The most engaged and scientifically literate users are likely to get the most out of this sort of tracking as they can bring wider knowledge/resources to bear to help them interpret their data — while a less informed user might take an overly simplistic read of what the information means.

There is also the risk that linking big bold stress triggers to food and other lifestyle events could lead to (or exacerbate) problems like eating disorders. The service wrapper and support are therefore a really key piece of making the most of what CGM tech can offer.

In short, poor UX decisions could have serious ramifications. And a lot of care and due diligence is needed over service design and delivery.

Longer term, having a snapshot view of blood glucose may — on its own — turn out to be far too limiting.

A more fully integrated tracking platform is likely to be needed to deliver the best understanding of an individual’s metabolism, drawing in a variety of signals and biomarkers. Although, right now, tracking glucose feels like a start; one which offers the chance to experiment with lifestyle tweaks that could accrue significant benefits over time — in a way that’s far more motivating than trying to figure out healthy individual dietary choices without any kind of real-time feedback.

Even just four weeks using the product yielded so many interesting tidbits and so much food for thought — avocado and egg is a super solid breakfast choice!; beer is a terrible spiker but natural cider looks (er) practically medicinal!; olives and nuts are truly the food of the Gods! — and the experience has led me to make some small but sustained lifestyle changes.

The jury is still out on whether those tweaks are genuinely worthwhile from a long-term health point of view. But given the changes weren’t especially radical, even if there’s only a tiny chance they have a benefit then, really, where’s the harm in that?

That said, another qualification: I do wonder whether (further) reducing the amount of carbs I eat — as a result of seeing how much they can spike me — might not have capped how much energy I have available for training purposes.

I already had a fairly low intake of carbs and it’s important to remember that food is also fuel — and energy needs vary. So a ‘spikes are bad, stability is best’ view on blood glucose may be too simplistic for an above average sporty lifestyle.

There is a real need to plug this data into relevant specialisms. A personal trainer would likely be able to make far more intelligent use of my results for me — based on knowing my individual fuel for training needs. Such a person may even be able to advise on dietary tweaks that could let me have my bread and eat it, so to speak.

But of course a personal trainer — or nutritionist — isn’t something everyone can afford or otherwise justify based on their (non-Olympic athlete) lifestyle. So on that front the product looks good value. (Even if you’re mostly getting raw data and need to do much of the wider interpretation yourself.)

How much does Ultrahuman’s Cyborg cost? The beta program is priced at ~$80 for two weeks (or $470 for 12 weeks). If you were paying a human personal trainer to be on your case and analyzing your data 24/7 it would be a lot more expensive than that — so it looks like pretty good value. (A decent personal trainer might cost $80 an hour.)

It’s important to emphasize that the app isn’t actively trying to be a full-time personal trainer. But it can do some basic things like offer exercise nudges if your blood glucose gets too high and — retrospectively — identify your “best workout zones”, aka optimal time windows to take exercise over the course a week based on how your body was fuelled. (“Do you see a trend? Use these times to your advantage to crush your next workout” was one suggestion it emailed me, although this nudge seemed more random than useful tbh.)

There are also a few (human) coaches on hand in the app to take questions and help you analyze your data. Plus you can always ask for help from other users via the invite-only Cyborg Slack channels. (Albeit, that’s crowdsourced wisdom, not dedicated professional support.) So the ‘relative value’ price-tag comes with the caveat that most of the time you’re on your own when it comes to drilling in and distilling more nuanced insights.

One more thought to ponder: As with every data-driven and ambitiously predictive AI product the Cyborg isn’t just training you; your data is training the Cyborg… So how much do you think 24/7 access to your biology is worth?

The value being derived from your highly intimate personal data flows two ways — and that upside isn’t necessarily being distributed equally. If you feel you’re getting enough value from the service that may not bother you. But privacy considerations are impossible to ignore.

Even if you’re comfortable sharing such intimate data with a commercial company in order to be able to access the service, Ultrahuman’s privacy policy for Cyborg notes some circumstances where your information may end up elsewhere — such as if it receives a subpoena it’s legally bound to respond to.

The policy also specifies that: “Anonymized, aggregated data may be shared with advertisers, research firms and other partners.” And robustly anonymizing health data has been shown to be notoriously difficult to do, even as the adtech industry has shown a rapacious appetite for triangulating and “sharing” data to better profile individuals for targeting — up to and including applying labels like “diabetes”. So your highly personal data leeching from a CGM into the modern web’s manipulative microtargeting ad matrix is not, unfortunately, impossible to imagine.

But let’s end on a personal note: What has Kumar himself learned from using CGM to track his glucose?

“For me the biggest learning has been around expanding my food spectrum and incorporating more of the foods that I like. Prior to this I pretty much followed a disciplined diet but couldn’t sustain it for long given it would affect social eating etc. With Cyborg, I’m able to understand how I can balance food with my activity. On the days I lift weights or am generally more active, I know that I have a little bit of extra flexibility around eating what I want to,” he tells TechCrunch.

“The other big learning — which is a work in progress — is around maintaining stable energy levels throughout the day. For me, stable glucose levels are broadly correlated to stable energy levels and this is what I’ve been trying to maintain during a particular week where I have a lot of reading work to do.”

---

 

Cutting edge competition

Here, past the quantified self trend’s needle-phobia line, is truly a wild(er) west — a lesser trodden arena of experimental startup opportunity. Naturally, it’s a lot more interesting than boring old step/sleep tracking, exactly because it’s so much less familiar.

There is a genuine sense of discovery as you fire the spring-loaded CGM sensor into your arm; feeling like a bit of a pioneer, involved in a kind of citizen science collective — with the fascinating opportunity to design and run experiments that interrogate the health of your own lifestyle.

On top of that, is the overarching possibility that what you learn personally might be useful to others — encouraged by Ultrahuman’s community-building efforts around Cyborg (such as its Slack channels, where early adopters are encouraged to share their learnings; as well as virtual and in person meet ups) — so there’s a ‘philanthropic mission’ feel as well.

Startups that are bold enough to get entangled in skin-puncturing machine-human interactions do have a chance to stand out. After all, mainstream tech giants simply can’t be that freaky. And it sets them apart from the wider wellness quantification crowd that’s plumped for a more quotidian biomarker to track.

That in turn means these startups have a chance to grab some very intimate biological data to feed their product dev, data science, AI models and algorithmic predictions; and — potentially — jockey themselves into position to race ahead as consumer appetite for personalized health services steps up.

On the blood glucose tracking front — an activity that has traditionally been associated with people who have conditions like diabetes (or pre-diabetes) — a large number of startups are now taking the plunge into willing recipients’ interstitial fluids.

As well as (India’s) Ultrahuman, with its still beta Cyborg service, there’s January AI, which does glucose tracking combined with heart rate monitor data to offer personalized food predictions and exercise ‘recipes’ to help you burn off any indulgent excess; Levels Heath, which has bagged backing from a16z; Signos, which is using CGMs to offer real-time weight loss advice; the athletic-performance focused Supersapiens; and NutriSense, which offers big picture soundbites around “optimizing” your “daily health performance”, to name a few of Ultrahuman’s US CGM-leveraging rivals.

There are more competitors in Europe — including (UK-based) Zoe, which is using data from large-scale microbiome studies to generate AI models to predict individual food responses. So as well as getting users to wear a blood glucose monitor, it also asks them to send in stool samples for lab analysis.

Also targeting glucose monitoring in the region is Germany’s Perfood (“personalized diet” for weight management); and Holland’s Clear Nutrition (“learn your unique responses to food” and “build your own nutrition plan”). At the time of writing, another European company in this nascent space, Finland’s Veristable — or Veri for short — was spotted advertising its 24/7 glucose monitor service on photo-sharing social network Instagram.

The ad pictured a very hipster look model sporting the same disc-shaped wearable which Ultrahuman’s service uses (and indeed many others do), taped over with a stylish grey patch vs the former’s black and white disk emblazoned with its tipped ‘K’ symbol. “End the Guessing Game of “What Should I Eat?” Veristable’s ad proclaimed, pointing Europeans toward a €159pm service.

Both Veri, which raised a seed round in June (per Crunchbase), and Ultrahuman — and a number of others — are using a CGM made by Abbott (the aforementioned FreeStyle Libre). This disc-shaped data-collecting device comes with a spring-loaded applicator that’s armed with a hollow needle. Once positioned in place you press down (firmly but not too firmly) and it fires the filament directly into your flesh.

The novelty here isn’t the tech itself — CGMs have been around for some years (Abbott’s FreeStyle Libre was introduced in 2016 for example, while Dexcom, another maker, got FDA approval for a fully interoperable CGM that could be used with other electronic diabetes management devices back in 2018) — it’s what they’re doing with it that’s experimental.

So while CGM tech has already been a transformative technology for people with diabetes and pre-diabetes, it’s only relatively recently there have been moves to commercialize it for a more general user who just wants to get to know their own body better.

Last summer, Dexcon gained FDA clearance for its real-time APIs for third party developers and devices. Fitness hardware maker Garmin was among the first wave of companies signed up to work with it to expand users’ access to their glucose data, albeit still with a focus on boosting utility for people with diabetes.

But investors have been quick to spot broader consumer potential — and are increasingly injecting funds to accelerate developments and get CGMs into many more arms.

Early last year, for example, January AI topped up with a further $8.8M; while Zoe bagged a $53M Series B in May 2021 (recently expanded when they added Balderton as an investor). Ultrahuman also announced a $17.5M in Series B (in August 2021); while Signos raised a $13M Series A in November.

As more data flows, it’s a safe bet that much more VC cash will follow. 

Ultrahuman’s PR points to the scale of the addressable potential market — talking about an emerging “metabolic health crisis”, and claiming that some 88%+ of Americans (and almost 80% of the global population) are “dealing with a metabolic disorder”; and thus could potentially benefit from joining its “Cyborg army”, as it brands early adopters.

So the potential addressable market is huge. Although any such wider onboarding looks like it will entail a steep learning curve — as startups seek to push beyond the low friction pond of early adopters and performance enthusiasts and step outside the quantified self and biohacking communities where this tech will naturally thrive.

A big part of Ultrahuman’s community-building efforts focus on encouraging users to share individual experiences and tips via invite-only Cyborg Slack channels and Townhalls, as well as signing up sporty influencers to evangelise the benefits of “performance fuelling” and other biohacking techniques that feed the purpose of sporting a CGM.

“The world currently has over 500M+ people who are diabetic but if you look at the problem holistically, you’d notice that there are almost 600M+ pre-diabetic people,” its PR goes on to claim, before suggesting the cure: CGM technology combined with “health score algorithms” and “instant health nudges” — which it argues “could help millions improve and help control / reverse this crisis”.

Whether millions of people can be sold on wearing a sensor in their skin remains to be seen.

But the technology may well evolve so it can be less invasive — and more mainstream-friendly — without losing too much accuracy. At which point there’s no reason to think it wouldn’t become a standard bit of fitness kit.

Whether a metabolic tracker subscription service is something millions of people will shell out for every month is another question. But once you’ve had a taster of this kind of data access it can be addictive. Even if you may also feel a bit ‘watched’ and judged as the sensor feeds back data on your lifestyle choices and the software scores how healthily you live.

It’s funny to imagine that the world’s unhealthy pursuit of something sweet to eat may, over time, get commercially rerouted into tracking and biohacking the rollercoaster ride of blood sugar — which should at least be a healthier fixation than blindly chasing the next sugar high.

 

Elvie, the women’s health tech pioneer behind a connected breast pump and smart pelvic floor exerciser, has topped up a Series C which it announced earlier this summer (July) — adding a further £12.7m to bring the total raised to £70 million ($97m).

The 2013-founded, UK-based startup previously raised a $42M Series B in 2019, and a $6M Series A in 2017 — when femtech startups were a lot rarer than they are now. Products designed for (and often by) women have gained a lot of momentum over this period as female-led startups have blazed a trail and shown there’s a sizeable market for femtech — leading investors to slow clock on to the opportunity too.

Analysts now project the femtech industry will become a $50 billion market by 2025.

Elvie says the Series C extension includes funds sponsored by the co-founders of Blume Equity – a PE firm that focuses on the food and health sectors – plus further capital from existing investors IPGL, Hiro Capital and Westerly Winds.

In July, when it announced the earlier ($80M) tranche of the raise, Elvie said the Series C was led by BGF and BlackRock alongside existing investors including Octopus Ventures.

The Series C will be used to drive for more growth through geographical expansion (including entering new markets) and diversifying its product portfolio to target other “key stages” in women’s lives, it said.

That means it’ll be splashing out on R&D to support product development — connected hardware that blends physical gadgetry with software still looks to be a strong focus — and also on strengthening its ops and infrastructure to prep for further scale.

Elvie sells four products at this stage: Its connected kegel trainer, and a wearable breast pump (plus two non-electric pumps).

Where the company goes next in terms of product will be an interesting one to watch.

Commenting in a statement, Tania Boler, CEO and founder, said: “Elvie is ready for the next phase of our growth. We have already revolutionized the categories we operate in, but we know that there is vast untapped potential to create better technology products and services for women in new areas.”

She added that Elvie’s goal is to create “the go-to destination for women’s health at all life stages” — selling “sophisticated, accurate and personalised solutions” to its target female consumer.

Fitness platform Ultrahuman has officially announced a $17.5 million Series B fund raise, with investment coming from early stage fund Alpha Wave Incubation, Steadview Capital, Nexus Venture Partners, Blume Ventures and Utsav Somani’s iSeed fund.

A number of founders and angel investors also participated in the Bangalore-headquartered startup’s Series B, including Tiger Global’s Scott Schleifer, Deepinder Goyal (CEO of Zomato), Kunal Shah (CEO of Cred), and Gaurav Munjal and Romain Saini (the CEO and co-founders of unacademy), among others. The latest tranche of funding brings its total raised to date to $25M.

While the subscription platform has been around since 2019, offering a fairly familiar blend of home workout videos, mindfulness content, sleep sessions and heart rate tracking (integrating with third party wearables like the Apple Watch), its latest fitness tool looks rather more novel — as it’s designed for monitoring metabolic activity by tracking the user’s glucose levels (aka, blood sugar).

Keeping tabs on blood sugar is essential for people living with diabetes. But in the US alone millions of people are prediabetic — meaning they have a higher than normal level of blood glucose and are at risk of developing diabetes, though they may not know it yet.

More broadly, Ultrahuman claims over a billion people in the world suffer from a metabolic health disorder — underlining the scale of the potential addressable market it’s eyeing. 

Having sustained high blood glucose is associated with multiple health issues so managing the condition with lifestyle changes like diet and exercise is advisable. Lifestyle changes can reduce elevated blood glucose and shrink or even avoid negative health impacts — such as by averting the risk of a prediabetic person going on to develop full blown diabetes.

But knowing what type of diet and exercise regime will work best for a particular person can be tricky — and involve a lot of frustrating trial and error — since people’s glucose responses to different food items can differ wildly.

These responses depend on a person’s metabolic health — which in turn depends on individual factors like microbiome diversity, stress levels, time of day, food ingredient and quality. (See also: Personalized nutrition startups like Zoe — which is similarly paying mind to blood glucose levels but as one component of a wider play to try to use big data and AI to decode the microbiome.) 

With metabolic health being so specific to each of us there’s a strong case for continuous glucose monitoring having widespread utility — certainly if the process and price-point can be made widely accessible.

Here, Ultrahuman is having a go at productizing the practice for a fitness enthusiast market — launching its first device in beta back in June — although the price-point it’s targeting is starting out fairly premium. 

The product (a wearable and a subscription service) — which it’s branded ‘Cyborg’ — consists of a skin patch that extracts glucose from the interstitial fluid under the skin, per founder and CEO, Mohit Kumar, with the data fed into a companion app for analysis and visualization.

The patch tracks the wearer’s blood glucose levels as they go about their day — eating, exercising, sleeping etc — with the biomarker used to trigger the app to nudge the user to “optimize your lifestyle”, as Ultrahuman’s website puts it — such as by alerting the user to a high blood glucose event and suggesting they take exercise to bring their level down.

If the product lives up to its promise of continuous glucose monitoring made easy, lovers of junk food could be in for a rude awakening as they’re served fast feedback on how their body copes (or, well, doesn’t) with their favorite snacks…

“We use medical grade sensors that have been used in the sports technology domain for the last 6-7 yrs with decent accuracy levels,” says Kumar when we ask about the specifics of the wearable technology it’s using. (The sensing hardware is being ‘worn’ here in the sense that it’s directly attached to (i.e. stuck into/on) bare skin.)

While Ultrahuman’s platform has plenty more vanilla fitness content, the company is now billing itself as a “metabolic fitness platform” — putting the nascent product front and center, even though the glucose tracking subscription service remains in closed beta for now.

The startup is operating a waitlist for sign-ups as it continues to hone the technology.   

Ultrahuman touts “thousands” of people signed up and waiting to get their hands on the glucose tracker service — and says it’s seeing 60% week over week growth in sign ups, with wider availability of the product slated for “early 2022”.

Some of the Series B cash will be used to make improvements to the quality of the glucose biomarkers ahead of a full product launch.

On the enhancements side, Kumar tells TechCrunch the team is exploring “other form factors and other types of sensors that could help us capture glucose in a more accurate way and for a longer duration than 14 days”, as they work to hone the wearable. (The current version of the skin-worn sensor only lasts two weeks before it must be replaced with another patch.)

“We want to add more biomarkers like HRV [heart-rate variability], sleep zones and respiratory rate to help people understand the impact of metabolic health on their recovery/sleep and vice-versa,” he adds.

Ultrahuman says it decided to focus on tracking glucose as its “main biomarker” as it can be used as a proxy for quantifying a number of fitness and wellness issues — making it a (potentially) very useful measure of individual health signals.

Or provided the startup’s technology is able to detect changes to glucose levels with enough sensitivity to be able to make meaningful recommendations per user.

“Glucose is interesting because it is a real-time biomarker that’s affected by exercise, sleep, stress and food,” says Kumar, adding: “We are able to help people make lifestyle changes across many vectors like nutrition, sleep, stress and exercise vs being unidimensional. It is also highly personalized as it guides you as per your body’s own response.”

He gives some examples of how the product could help users by identifying beneficial tweaks they could make to their diet and exercise regimes — such as figuring out which foods in their current diet yield “a healthy metabolic response” vs those that “need more optimization” (aka, avoiding the dreaded sugar crash). Or by helping users identify “a great meal window” for their lifestyle — based in their body’s glucose consumption rate.

Other helpful nudges he suggests the service can provide to sensor-wearing users — with an eye on athletes and fitness fanatics — is how best to fuel up before exercise to perform optimally.

Optimizing the last meal of the day to improve sleep efficiency is another suggestion.

If Ultrahuman’s Cyborg can do all that with a (bearably) wearable skin patch and a bit of clever algorithmic analysis it could take the quantified self trend to the next level.

A simple stick-on sensor-plus-app that passively amplifies internal biological signals and translates individual biomarkers into highly actionable real-time personalized health insights could be the start of something huge in preventative healthcare.

Again, though, Ultrahuman’s early pricing suggests there will be some fairly hard limits on who is able to tap in here.

Early adopters in the closed beta are shelling out $80 per month for the subscription service, per Kumar. And — at least for now — the startup is eyeing adding more bells and whistles, rather than fewer. “[Product pricing] will mostly be in the same range but may introduce more services/premium features on top of this,” he confirms.

The (typically higher) cost of eating healthily and having enough leisure time to be able to look after your body by taking exercise are other hard socioeconomic limits that won’t be fixed by a wearable, no matter how smart.

 

Personalized nutrition startup Zoe — named not for a person but after the Greek word for ‘life’ — has topped up its Series B round with $20M, bringing the total raised to $53M.

The latest close of the B round was led by Ahren Innovation Capital, which the startup notes counts two Nobel laureates as science partners. Also participating are two former American football players, Eli Manning and Ositadimma “Osi” Umenyiora; Boston, US-based seed fund Accomplice; healthcare-focused VC firm THVC and early stage European VC, Daphni.

The U.K.- and U.S.-based startup was founded back in 2017 but operated in stealth mode for three years, while it was conducting research into the microbiome — working with scientists from Massachusetts General Hospital, Stanford Medicine, Harvard T.H. Chan School of Public Health, and King’s College London.

One of the founders, professor Tim Spector of King’s College — who is also the author of a number of popular science books focused on food — became interested in the role of food (generally) and the microbiome (in particular) on overall health after spending decades researching twins to try to understand the role of genetics (nature) vs nurture (environmental and lifestyle factors) on human health.

Zoe used data from two large-scale microbiome studies to build its first algorithm which it began commercializing last September — launching its first product into the U.S. market: A home testing kit that enables program participants to learn how their body responds to different foods and get personalized nutrition advice.

The program costs around $360 (which Zoe takes in six instalments) and requires participants to (self) administer a number of tests so that it can analyze their biology, gleaning information about their metabolic and gut health by looking at changes in blood lipids, blood sugar levels and the types of bacteria in their gut.

Zoe uses big data and machine learning to come up with predictive insights on how people will respond to different foods so that it can offer individuals guided advice on what and how to eat, with the goal of improving gut health and reducing inflammatory responses caused by diet.

The combination of biological responses it analyzes sets it apart from other personalized nutrition startups with products focused on measuring one element (such as blood sugar) — is the claim.

But, to be clear, Zoe’s first product is not a regulated medical device — and its FAQ clearly states that it does not offer medical diagnosis or treatment for specific conditions. Instead it says only that it’s “a tool that is meant for general wellness purposes only”. So — for now — users have to take it on trust that the nutrition advice it dishes up is actually helpful for them.

The field of scientific research into the microbiome is undoubtedly early — Zoe’s co-founder states that very clearly when we talk — so there’s a strong component here, as is often the case when startups seek to use data and AI to generate valuable personalized predictions, whereby early adopters are helping to further Zoe’s research by contributing their data. Potentially ahead of the sought for individual efficacy, given so much is still unknown around how what we eat affects our health.

For those willing to take a punt (and pay up), they get an individual report detailing their biological responses to specific foods that compares them to thousands of others. The startup also provides them with individualized ‘Zoe’ scores for specific foods in order to support meal planning that’s touted as healthier for them.

“Reduce your dietary inflammation and improve gut health with a 4 week plan tailored to your unique biology and life,” runs the blurb on Zoe’s website. “Built around your food scores, our app will teach you how to make smart swaps, week by week.”

The marketing also claims no food is “off limits” — implying there’s a difference between Zoe’s custom food scores and (weight-loss focused) diets that perhaps require people to cut out a food group (or groups) entirely.

“Our aim is to empower you with the information and tools you need to make the best decisions for your body,” is Zoe’s smooth claim.

The underlying premise is that each person’s biology responds differently to different foods. Or, to put it another way, while we all most likely know at least one person who stays rake-thin and (seemingly) healthy regardless of what (or even how much) they eat, if we ate the same diet we’d probably expect much less pleasing results.

“What we’re able to start scientifically putting some evidence behind is something that people have talked about for a long time,” says co-founder George Hadjigeorgiou. “It’s early [for scientific research into the microbiome] but we have shown now to the world that even twins have different gut microbiomes, we can change our gut microbiomes through diet, lifestyle and how we live — and also that there are associations around particular [gut] bacteria and foods and a way to improve them which people can actually do through our product.”

Users of Zoe’s first product need to be willing (and able) to get pretty involved with their own biology — collecting stool samples, performing finger prick tests and wearing a blood glucose monitor to feed in data so it can analyze how their body responds to different foods and offer up personalized nutrition advice.

Another component of its study of biological responses to food has involved thousands of people eating “special scientific muffins”, which it makes to standardized recipes, so it can benchmark and compare nutritional responses to a particular blend of calories, carbohydrate, fat, and protein.

While eating muffins for science sounds pretty fine, the level of intervention required to make use of Zoe’s first at-home test kit product is unlikely to appeal to those with only a casual interest in improving their nutrition.

Hadjigeorgiou readily agrees the program, as it is now, is for those with a particular problem to solve that can be linked to diet/nutrition (whether obesity, high cholesterol or a disease like type 2 diabetes, and so on). But he says Zoe’s goal is to be able to open up access to personalized nutrition advice much more widely as it keeps gathering more data and insights.

“The idea is, as always, we start with a focused set of people with problems to solve who we believe will have a life-changing experience,” he tells TechCrunch. “At this point we are not trying to create a product for everyone — and we understand that that has limitations in terms of how much we scale in the beginning. Although even still within this focused group of people I can assure you there’s tonnes of people!

“But absolutely the whole idea is that after we get a first [set of users]… then with more data and with more experience we can simplify and start making this simpler and more accessible — both in terms of its simplicity and also it’s price. So more and more people. Because at the end of the day everyone has this right to be able to optimize and understand and be in control — and we want to make that available to everyone.

“Regardless of background and regardless of socio-economic status. And, in fact, many of the people who have the biggest problems around health etc are the ones who have maybe less means and ability to do that.”

Zoe isn’t disclosing how many early users it’s onboarded so far but Hadjigeorgiou says demand is high (it’s currently operating a wait-list for new sign ups).

He also touts promising early results from interim trial with its first users — saying participants experienced more energy (90%), felt less hunger (80%) and lost an average of 11 pounds after three months of following their AI-aided, personalized nutrition plan. Albeit, without data on how many people are involved in the trials it’s not possible to quantify the value of those metrics.

The extra Series B funding will be used to accelerate the rollout of availability of the program, with a U.K. launch planned for this year — and other geographies on the cards for 2022. Spending will also go on continued recruitment in engineering and science, it says.

Zoe already grabbed some eyeballs last year, as the coronavirus pandemic hit the West, when it launched a COVID-19 symptom self-reporting app. It has used that data to help scientists and policy makers understand how the virus affects people.

The Zoe COVID-19 app has had some 5M users over the last year, per Hadjigeorgiou — who points to that (not-for-profit) effort as an example of the kind of transformative intervention the company hopes to drive in the nutrition space down the line.

“Overnight we got millions and millions of people contributing to help uncover new insights around science around COVID-19,” he says, highlighting that it’s been able to publish a number of research papers based on data contributed by app users. “For example the lack of smell and taste… was something that we first [were able to prove] scientifically, and then it became — because of that — an official symptom in the list of the government in the U.K.

“So that was a great example how through the participation of people — in a very, very fast way, which we couldn’t predict when we launched it — we managed to have a big impact.”

Returning to diet, aren’t there some pretty simple ‘rules of thumb’ that anyone can apply to eat more healthily — i.e. without the need to shell out for a bespoke nutrition plan? Basic stuff like eat your greens, avoid processed foods and cut down (or out) sugar?

“There are definitely rules of thumb,” Hadjigeorgiou agrees. “We’ll be crazy to say they’re not. I think it all comes back to the point that although there are rules of thumb and over time — and also through our research, for example — they can become better, the fact of the matter is that most people are becoming less and less healthy. And the fact of the matter is that life is messy and people do not eat even according to these rules of thumb so I think part of the challenge is… [to] educate and empower people for their messy lives and their lifestyle to actually make better choices and apply them in a way that’s sustainable and motivating so they can be healthier.

“And that’s what we’re finding with our customers. We are helping them to make these choices in an empowering way — they don’t need to count calories, they don’t need to restrict themselves through a Keto [diet] regime or something like that. We basically empower them to understand this is the impact food has on your body — real time, how your blood sugar levels change, how your bacteria change, how your blood fat levels changes. And through that empowerment through insight then we say hey, now we’ll give you this course, it’s very simple, it’s like a game — and we’ll given you all these tools to combine different foods, make foods work for you. No food is off limits — but try to eat most days a 75 score [based on the food points Zoe’s app assigns].

“In that very empowering way we see people get very excited, they see a fun game that is also impacting their gut and metabolism and they start feeling these amazing effects — in terms of less hunger, more energy, losing weight and over time as well evolving their health. That’s why they say it’s life changing as well.”

Gamifying research for the goal of a greater good? To the average person that surely sounds more appetitizing than ‘eat your greens’.

Though, as Hadjigeorgiou concedes, research in the field of microbiome — where Zoe’s commercial interests and research USP lie — is “early”. Which means that gathering more data to do more research will remain a key component of the business for the foreseeable future. And with so much still to be understood about the complex interactions between food, exercise and other lifestyle factors and human health, the mission is indeed massive.

In the meanwhile, Zoe will be taking it one suggestive nudge at a time.

“Sugar is bad, kale’s great but the whole kind of magic happens in the middle,” Hadjigeorgiou goes on. “Is oatmeal good for you? Is rice good for you? Is wholewheat pasta good for you? How do you combine wholewheat pasta and butter? How much do you have? This is where basically most of our life happens.

“Because people don’t eat ice-cream the whole day and people don’t eat kale the whole day. They eat all these other foods in the middle and that’s where the magic is — knowing how much to have, how to combine them to make it better, how to combine it with exercise to make it better? How to eat a food that doesn’t dip your sugar levels three hours after you eat it which causes hunger for you. Theses are all the things we’re able to predict and present in a simple and compelling way through a score system to people — and in turn help them [understand their] metabolic response to food.”

Human rights NGO, Amnesty International, has written to the EU’s competition regulator calling for Google’s acquisition of wearable maker Fitbit to be blocked — unless meaningful safeguards can be baked in.

The tech giant announced its intent to splash $2.1BN to acquire Fitbit a year ago but has yet to gain regulatory approval for the deal in the European Union.

In a letter addressed to the blocs competition chief, Margrethe Vestager, Amnesty writes: “The Commission must ensure that the merger does not proceed unless the two business enterprises can demonstrate that they have taken adequate account of the human rights risks and implemented strong and meaningful safeguards that prevent and mitigate these risks in the future.”

The letter urges the Commission to take heed of an earlier call by a coalition of civil society groups also raising concerns about the merger for “minimum remedies” which regulators must guarantee before any approval.

In a report last year the NGO attacked the business model of Google and Facebook — arguing that the “surveillance giants” enable human rights harm “at a population scale”.

Amnesty warns now that Google is “incentivized to merge and aggregate data across its different platforms” as a consequence of that surveillance-based business model.

“Google’s business model incentivizes the company to continuously seek more data on more people across the online world and into the physical world. The merger with Fitbit is a clear example of this expansionist approach to data extraction, enabling the company to extend its data collection into the health and wearables sector,” it writes. “The sheer scale of the intrusion of Google’s business model into our private lives is an unprecedented interference with our privacy, and in fact has undermined the very essence of privacy.”

We’ve reached out to the Commission and Google for a response to Amnesty’s letter.

Google’s plan to gobble Fitbit and its health tracking data has been stalled as EU regulators dig into competition concerns. Vestager elected to open an in-depth probe in August, saying she wanted to make sure the deal wouldn’t distort competition by further entrenching Google’s dominance of the online ad market.

The Commission has also voiced concerns about the risk of Google locking other wearable device makers out of its Android mobile ecosystem.

However concerns over Google’s plan to gobble up Fitbit range wider than the risk of it getting more market muscle if the deal gets waved through.

Put simply, letting sensitive health data fall into the hands of an advertising giant is a privacy trashfire.

Amnesty International is just the latest rights watcher to call for the merger to be blocked. Privacy campaign groups and the EU’s own data protection advisor have been warning for months against letting the tech giant gobble up sensitive health data.

The Commission’s decision to scrutinize the acquisition rather than waiving it through with a cursory look has led Google to make a number of concessions in an attempt to get it cleared — including a pledge not to use Fitbit data for ad targeting and to guarantee support for other wearables makers to operate on Android.

In its letter, Amnesty argues that the ‘safeguards’ Google has offered are not enough.

“The company’s past practice around privacy further heighten the need for strict safeguards,” it warns, pointing to examples such as Google combining data from advertising network DoubleClick after it had acquired that business with personal data collected from its other platforms.

“The European Data Protection Board has recognized the risks of the merger, stating that the “combination and accumulation of sensitive personal data” by Google could entail a “high level of risk” to the rights to privacy and data protection,” it adds.

As well as undermining people’s privacy, Google’s use of algorithms fed with personal data to generate profiles of Internet users in order to predict their behavior erodes what Amnesty describes as “the critical principle that all people should enjoy equal access to their human rights”.

“This risk is heightened when profiling is deployed in contexts that touch directly on people’s economic, social and cultural rights, such as the right to health where people may suffer unequal treatment based on predictions about their health, and as such must be taken into account in the context of health and fitness data,” it suggests.

“This power of the platforms has not only exacerbated and magnified their rights impacts but has also created a situation in which it is very difficult to hold the companies to account, or for those affected to access an effective remedy,” Amnesty adds, noting that while big tech companies have faced a number of regulatory actions around the world none has so far been able to derail what it calls “the fundamental drivers of the surveillance-based business model”.

So far the Commission has stood firm in taking its time to consider the issue in detail.

A series of extensions mean a decision on whether to allow the Google-Fitbit merger may not come until early 2021. Though we understand the bloc’s national competition authorities are meeting to discuss the merger at the start of December so it’s possible a decision could be issued before the end of the year.

Per EU merger law, the Commission college takes the final decision — with a requirement to take “utmost account” of the opinion of the Member States’ advisory committee (though it’s not legally binding).

So it’s ultimately up to Brussels to determine whether Google-Fitbit gets green lit.

In recent years, competition chief Vestager, who is also EVP for the Commission’s digital strategy, has said she favors tighter regulation as a tool for ensuring businesses comply with the EU’s rules, rather than blocking market access or outright bans on certain practices.

She has also voiced opposition to breaking up tech giants, again preferring to advocate for imposing controls on how they can use data as a way to rebalance digital markets.

To date, the Commission has never blocked a tech merger. Though it has had its fingers burnt by big tech’s misleading filings — so has its own reputation to consider above reaching for the usual rubberstamp.

Simultaneously, EU lawmakers are working on a proposal for an ex ante regulation to address competition concerns in digital markets that would put specific rules and obligations on dominant players like Google — again in areas such as data use and data access.

That plan is due to be presented early next month — so it’s another factor which may be adding to the delay to the Commission’s Google-Fitbit decision.

Fertility tracking has seen an explosion of startup activity in recent years. Femtech startup Lady Technologies is adding to this rich mix with the full U.S. launch of a dual-purpose device, called kegg, that’s designed to measure hormonal changes in a woman’s cervical fluid to help her determine the chance of conception on a given day.

The egg-shaped gizmo, which features a gold-plated steel cap and band ringing its tip, as well as a silicone tail to house its Bluetooth radio (so it can chat to the companion app), doubles as a connected pelvic floor trainer (the ‘k’ in kegg is for ‘kegels’) — taking a leaf out of UK femtech pioneer Elvie’s playbook. Though the two-in-one function is a new twist.

Kegg relies on a technology called impedance to sense electrolyte levels in a woman’s cervical fluid in order to detect the hormonal switch from estrogen to progesterone dominance that accompanies ovulation — via a daily test that’s touted as taking just two minutes. (If you’re also using it for the optional kegal exercises that would take a bit longer.)

“A minute electrical impulse at a specific frequency is emitted from the gold plated electrodes on the kegg and received by the other (this process is then reversed). By sensing the changing trends in the impedance, we’re able to detect the hormonal change and make a prediction to the user,” explains CEO and founder Kristina Cahojova. “Since every woman’s fluids are slightly different, kegg needs to record at least one fertile window to provide personalized predictions.”

“We have numerous patents on the underlying design of kegg and key aspects of how it operates,” she adds.

Kegg was unveiled on the TechCrunch Disrupt SF stage, back in 2018, as part of our startup battlefield competition (though it didn’t go on to win). Fast forward two years and it’s now officially launching out of beta to offer the FDA-registered gizmo to the U.S. market — priced at $275.

It’s announcing a $1.5M seed round too, with investors including Crescent Ridge Partners, SOSV, Texas Halo Fund, Fermata Fund and MegaForce, as well as some unnamed angel investors.

Commenting in a statement, Samina Hydery, kegg advisor and women’s health investor, said: “Investor interest in femtech and fertility has accelerated over the last few years. While I’ve seen an influx of ovulation prediction kits, at-home blood tests, menstrual tracking apps, and temperature monitors in the consumer market, kegg’s value proposition became clear once I spoke with women about their experiences trying to conceive and medical researchers in the field. It’s hard not to get excited by the various growth vectors that can expand kegg’s market in the future — from being used as a tool for natural family planning to helping monitor postpartum/perimenopausal health.”

“We pride ourselves in having almost half of our investors women,” notes Cahojova — whose inspiration for building kegg was personal; having suffered from irregular menstrual cycles herself.

“I didn’t want to be treated with hormones. When I talked to fertility instructors or a specialized fertility doctor, all they wanted to know about was my patterns of cervical fluid. Why? Because the fertile window is defined only by the presence of fertile cervical fluid, having a positive LH [luteinizing hormone] test is nice but it won’t help you get information to fix your cycles. That’s why so many fertility doctors are interested in cervical fluid and that is why so many women are told to track it with their fingers,” she explains.

“How on earth are you supposed to be able to track objectively something so important, yet, private without the help of technology? I was frustrated and angry that every company that I talked to didn’t have a solution and didn’t want to make this so needed product because it ‘would have to go into the vagina’. So I set out to make a product that would help me and women like me.”

Thus far kegg has been hitting a chord with U.S. women of reproductive age who are trying for a baby, according to Cahojova — who says her startup has built a 2,000-strong community of fertility-tracking women over kegg’s beta period.

“Our typical user is a woman in her reproductive age,” she says. “Our users are in long-term relationships or married and they likely have been actively trying to conceive for more than three months. Fifty percent are trying to conceive their first child, while the remaining are already mothers.

“Our customers have experience with BBT (body basal temperature charting) or LH tests (ovulation tests) and they are overall interested in holistic fertility and wellness, not in medication. They also prefer the convenience of kegg over other methods that either need to be worn throughout the night or used more frequently.”

“Each woman is unique and so are her cycles,” she adds. “Unlike ovulation trackers, kegg helps women understand their fertile window and cyclical fertility and follow their own patterns. Usually women take up to six months to learn how to read cervical fluid patterns. Our customers report that kegg gives them confidence and they feel empowered. Many keggsters conceived with kegg after years of trying because kegg gave them trends beyond ovulation. Nothing makes me more happy than an email from a customer whose life changed thanks to my work and kegg.” (On that it says “several” women have reported successful pregnancies using kegg since the beta launch in 2018.)

The startup also has its eye on international expansion, including to Asia (with the support of its Japanese-market focused investor Fermata) — with a plan to launch kegg in Singapore in late October, and in Japan and Canada next year.

While the kegg has a core focus on fertility tracking (and a secondary feature as a connected pelvic floor trainer), Cahojova is excited about wider possibilities for women’s health that she hopes will be opened up as they’re able to take in and crunch more data.

Kegg users’ impedance readings are uploaded to the startup’s cloud for analysis, so its algorithms can make a personalized fertility prediction. But its website also notes it uses ‘anonymized/pseudonymized’ data for research into women’s health. (Cahojova specifies users’ personal data is never shared outside the company. “Any data we offer to researchers we work with is completed anonymized,” is her privacy promise.)

Asked what areas of research she’s hoping kegg will help advance, she tells us: “Researchers have noted that health issues can affect typical electrolyte cycles. In many of our internal studies we’ve seen examples where readings were ‘out of norm’ for the user. In case after case we found evidence of underlying health issues (for example infections) were the cause. In the future our goal is to understand how kegg can help monitor overall cervical health.”

Cahojova also says the device is being used by fertility instructors and doctors to help with monitoring their patients. “The beauty of kegg is that by having a user friendly and modern device that women like to use we can get data on changes of vaginal fluids on a large scale. With kegg data we also hope to help doctors finally answer their billion dollar question — how can they improve the quality of cervical fluid.”

“We are supportive of science and are open for research collaborations,” she adds. “We provided kegg for independent peer-reviewed clinical study under Dr.Gabriela López Armas, MD, PhD, for her research on kegg and other fertility trackers. All the participants finished the protocols in summer of 2020 and the study is to be published independently in the near future.”

While the business model for kegg is currently fixed price hardware sales, Cahojova says the startup is looking at offering subscription packages in future. “In the future, we want to offer more to our users, e.g.: connecting them to specialists to review their cycles or view of additional layers of information. Once we have enhanced services ready, we’ll look at switching to a subscription model,” she adds.

It has been over four years since Project Jacquard, Google’s smart fabric technology, made its debut at the I/O developer conference. Launched out of what was then Google’s ATAP unit, Jacquard is currently best known for being available on Levi’s jeans jackets, but Saint Laurent also launched its $1,000 Cit-e Backpack with built-in Jacquard technology. Today, Google is adding a fourth product to the Jacquard lineup with the launch of the Samsonite Konnect-i backpacker, which, at $200 for the Slim version and $220 for the Standard edition, is a bit more friendly on the wallet than the Saint Laurent backpack.

Jacquard, in case you need a refresher, is Google’s technology for adding touch sensitivity to fabrics. That means you can touch the sleeve of your jacket or, in this case, the strap of your backpack, to trigger a handful of functions on your phone. The whole system is powered by a small tag (which you charge via a mini-USB port). That tag can also relay notifications through its built-in LED and a small vibration motor.

The number of gestures — and what they can trigger — is relatively limited, especially since you can only really assign three gestures: brush up, brush down and double-tap. You can assign standard media controls to these (think brush up for “next song”), drop a pin to save a place, hear the current time, ping your phone, hear directions to your next waypoint or arrival time or trigger the Google Assistant. Gestures can also trigger your phone’s shutter to take a selfie and there’s a “light” function that lights up the Jacquard tag’s LED. Why this last function exists isn’t quite clear to me because that LED is weak. Google says it can help you get noticed in a crowd or stay visible at night, but unless you’re trying to be found in the darkest of caves, nobody will be able to see it.

As you can see, the main idea here is to let you access some of your phone’s functions while walking through the city with your headphones on.

It’s been about a year since Google and Levi’s launched the Jacquard-enabled trucker jacket. At the time, that was the launch of Jacquard 2.0, with a couple of additional features and a new dongle that now works across products. At the time, our review and those from our peers were pretty tepid. I’m not sure it’ll be all that different this time around.

I’ve tried out the backpack for the last few days. Like before, Jacquard does what it promises to do. The gesture recognition worked as expected. Alerts from my phone made the tag vibrate and the backpack itself is comfortable, if not the flashiest entry into the market. It’s a Samsonite, though, and the target market here isn’t necessarily college students but business travelers (though that market is pretty dead for the time being).

The backpack itself comes in two versions: slim and standard. The only real difference here is that the slim version has a vertical zipper and the standard version a horizontal one. It features plenty of pockets, a padded laptop compartment and everything else you’d want from a modern backpack. I could easily see myself going on a business trip with it.

Like before, the question remains whether Jacquard is a gimmick or actually a useful technology. Thanks to the pandemic, most of us aren’t heading out as much as we used to — and we’re definitely not going on a lot of trips. Maybe it’s not the right product for this time, but I can see myself using it more than the jacket once all of this is over. Chances are I’ll use a backpack wherever I go, after all, whereas I don’t wear a jacket half the year.  The promise of Jacquard is to allow you to focus on the world around you, without the distractions of your phone. For that to work, it needs to be ubiquitous or you’ll just forget you ever had it. That works better on a backpack than a jacket — at least for me.

Whether that’s worth $200 to you is a decision you must make for yourself.

Google has offered a second round of concessions to try to persuade European regulators to clear its acquisition of wearables maker Fitbit .

The deal has been stalled by concerns over its impact on consumer privacy and competition in the wearables market.

Last week the deadline for EU regulators to take a decision was extended for another couple of weeks — potentially pushing it out to almost the end of the year.

However a report by Reuters today claims the acquisition is set to be greenlit after the latest round of ‘commitments’ from Google — with the news agency citing ‘people familiar with the matter’.

The European Commission declined to comment on the report.

Google confirmed it has sent a new set of commitments to the European Commission — reiterating an earlier pledge not to use Fitbit health and wellness data for advertising, which it said it has now strengthened by providing for additional monitoring of the data separation requirements. 

It also said it’s committing to support third-party wearable manufacturers as part of the Android ecosystem (via Android APIs for wearable devices), and maintain third-parties’ existing access to Fitbit users’ data via APIs with user consent. 

“This deal is about devices, not data. The wearables space is highly crowded, and we believe the combination of Google and Fitbit’s hardware efforts will increase competition in the sector, benefiting consumers and making the next generation of devices better and more affordable,” a Google spokesperson said in a statement.

“We have been working with the European Commission on an updated approach to safeguard consumers’ expectations that Fitbit device data won’t be used for advertising.  We’re also formalizing our longstanding commitment to supporting other wearable manufacturers on Android and to continue to allow Fitbit users to connect to third party services via APIs if they want to.”

European antitrust regulators now have until almost the end of the year to take a decision on whether to green light Google’s planned acquisition of Fitbit.

The tech giant announced its intention to buy the fitness tracking wearable maker in November 2019, saying it would shell out $2.1 billion in cash to make off with Fitbit and the health data it holds on some 28M+ users.

EU regulators were quick to sound the alarm about letting the tech giant go shopping for such a major cache of sensitive personal data, with the European Data Protection Board warned in February that the proposed purchase poses a huge risk to privacy.

There is also a parallel concern that Fitbit’s fitness data could further consolidate Google’s regional dominance in the ad market. And last month EU competition regulators announced a full antitrust probe — saying then they would take a decision within 90 working days. That deadline has now been extended by a further two weeks.

A Commission spokeswoman confirmed the earlier provisional December 9 deadline has been pushed on “in agreement with the parties” — citing Article 10(3) of the EU’s Merger Regulation.

“The provisional legal deadline for a final decision in this case is now December 23, 2020,” she added.

The Commission has not offered any detail on the reason for allocating more time to take a decision.

When EU regulators announced the in-depth probe, the Commission said it was concerned data gathered by Fitbit could lead to a distortion of competition if Google was allowed to assimilate the wearable maker and “further entrench” its dominance in online ad markets.

Other concerns include the impact on the nascent digital healthcare sector, and whether Google might be incentivised to degrade the interoperability of rival wearables with its Android OS once it has its own hardware skin in the game.

The tech giant, meanwhile, has offered assurances around the deal in an attempt to get it cleared — claiming ahead of the Commission’s probe announcement it would not use Fitbit health data for ad targeting, and suggesting that it would create a ‘data silo’ for Fitbit data to keep it separate from other data holdings.

However regulators have expressed scepticism — with the Commission writing last month that the “data silo commitment proposed by Google is insufficient to clearly dismiss the serious doubts identified at this stage as to the effects of the transaction”.

It remains to be seen what the bloc’s competition regulators conclude after taking a longer and harder look at the deal — and it’s worth noting they are simultaneously consulting on whether to give themselves new powers to be able to intervene faster to regulate digital markets — but Google’s hopes of friction-free regulatory clearance and being able to hit the ground running in 2020 with Fitbit’s data in its pocket have certainly not come to pass.