a16z's State of Longevity: 83% of Wearable Users Check Daily, and 55% Have Never Acted on Their Data

Most people who wear a health tracker look at it several times a day. Over a year that is more than 700 glances at a sleep score, a heart rate or a step count. And according to a survey published this week by Andreessen Horowitz, for most of them not one of those glances has ever changed what they did.

The number sits in a16z’s State of Longevity report, published on 3 November by Daisy Wolf and Eva Steinman [1]. Of the wearable users surveyed, 83% check their device every day. 55% say its data has never prompted a behaviour change or a doctor’s visit.

The report covers a lot of ground: peptides, cholesterol pills, microplastics, brain-computer interfaces. But for anyone building with health data, that gap is the finding that matters, and the report’s own thesis explains why. As AI makes answers cheap, a16z argues, value moves to whoever can turn them into action [1]. This is what the report says, what its survey found, and why so little wearable data becomes action.

What is a16z’s State of Longevity report?

A survey of 500 randomly selected US consumers and six trends that a16z’s Bio + Health team believes are reshaping consumer health [1]. Health and longevity, the authors argue, have moved from biohacker hobby to mainstream behaviour, pushed along by COVID-19 and by AI tools that made medical information easy to ask about.

Three of the six trends, the healthcare home, wearables and the data behind GLP-1 and cholesterol care, are about the same thing: more health data about each person, from more places, with more expected of it.

What did the survey find about wearables?

That people are deeply engaged with their devices and rarely moved by them. Of the 500 adults surveyed, 54% use a wearable, 270 people [1]. Among them:

• 60.2% check it several times a day and 23.3% once a day. Only 3.4% almost never look.
• 45% say their wearable data has ever prompted a behaviour change or a doctor’s visit. 55% say it never has.
• 71.9% would wear a second device if it measured something new.

The bar for “acted” is low. The question was whether the data had ever prompted a change or a visit, not whether it does so regularly. A user who once went to bed earlier after a poor sleep score counts. More than half still said no.

And the appetite for more is striking: most of those users want a second device. More measurement is not what is missing.

Why doesn’t wearable data lead to action?

Because a lot has to go right between a reading and a change, and every step leaks. We have spent the past few weeks measuring those steps one at a time. Put together, they explain the 55%.

1. The data arrives late. iOS wakes an app for new step data at most once an hour, and not reliably [2]. A message that says “you’ve barely moved today” at 3 pm may be describing 11 am. We covered every source of delay in why wearable data arrives late.

2. The comparison cries wolf. An alert is only as good as the baseline it compares against. In our simulation, a 7-day baseline flagged about 1 normal day in 12 as unusual [3]. Users learn to ignore an alert faster than they learn to trust it, which is why we recommend 14 to 28 days in how many days a wearable baseline needs.

3. The prompt moves the hour, not the day. In a randomised trial, a notification made people 3.5 times as likely to open a health app in the next hour, but only 1.3 times as likely over the next 24 hours, and a month of daily reminders did not keep users significantly longer [4]. More in do health app push notifications work.

4. The prompt wears off. In the HeartSteps trial, walking suggestions more than doubled steps in the following half hour at first. By the fourth week the effect was gone, and participants said the suggestions had become boring [5].

5. The change fades. Across 39 systematic reviews, activity trackers added about 1,800 steps a day at first and about 500 over the long term [6]. That is real and worth having, and most of the early gain does not last. More in do fitness trackers work.

There is a sixth leak that the survey hints at: most products stop at the number. A sleep score of 62 is information. Whether to go to bed earlier, skip a hard workout, or book a check-up is a decision, and most apps leave the user to make it alone. That is the gap a16z calls the action layer.

What is a healthcare home, and who is building it?

The single app a person turns to first for anything health related: one that holds their records, lab results, imaging and wearable data, explains them, and arranges the next step, whether that is a test, an appointment, a prescription or follow-up care [1]. a16z’s point is that the hard part is not the answer. Asking AI a health question is already easy. Booking the follow-up is not.

Demand is there, with a catch. 71.4% of respondents said they would use such an app; 49.2% said they would pay for it [1]. The 22-point gap between using and paying is the value a healthcare home still has to prove.

The contenders a16z names come at it from different starting points [1]:

• Labs and imaging: Function Health
• Daily habits: Oura and WHOOP
• Care delivery: General Medicine, Counsel and Ferry Health
• Large existing audiences: OpenAI, Anthropic, Amazon and Apple

a16z expects room for several winners. Whoever wins inherits the same problem underneath. A healthcare home is only as good as the data it can reach. Its users wear different brands, and nearly three in four would add a second device [1], so the same night’s sleep or the same day’s steps will arrive twice, from two sources, with different numbers. Reconciling that, as we set out in 12 checks for wrong wearable data, has to happen before any AI explains anything. And explanation itself is less solved than it looks: in a 2026 benchmark on real users’ wearable data, the best AI model answered 73% of questions correctly, as we covered in how accurate AI health coaches are.

What is the third generation of wearables?

a16z frames wearables in three acts [1]. Act 1 was counting steps, led by Fitbit. Act 2 was sleep, heart rate and recovery, led by Oura, WHOOP and Apple Watch. Act 3 is devices built to answer specific health questions: estimating hormone patterns without a blood draw, tracking blood flow from an earring, flagging night-time patterns linked to blood pressure, logging GLP-1 doses alongside the data they affect. The long-term ambition is a non-invasive device that tracks many biomarkers the way a continuous glucose monitor tracks one.

Regulation has opened the door, within limits. In January 2026 the FDA updated its general wellness guidance to clarify when non-invasive wearables can offer wellness insights without medical-device clearance [7]. In June it closed its warning letter to WHOOP after the company changed its Blood Pressure Insights feature and labelling [8], which we covered in cuffless blood pressure wearables. a16z’s conclusion is that there is more room to build, provided products stay within wellness claims [1].

For anyone building on top of wearables, Act 3 means more of everything: more devices per person, more metrics with no shared definition, and more features whose wording decides whether they are wellness or medicine.

What should health apps, insurers and wellness programmes do now?

1. Measure action, not attention. Daily checking is already high. The metric that separates products is the share of users who did something because of the data. We set out how to measure it in the health app metrics that predict retention.
2. Ship every insight with one next step. Not “your HRV is down 18%” but what to do about it today: an easier workout, an earlier night, a follow-up if the change persists.
3. Earn the right to alert. Use a 14 to 28-day baseline, require a change to persist before flagging it, and say how much data a comparison rests on.
4. Prompt sparingly and at the right moment. One or two prompts a day per behaviour, timed to the person’s routine, and none when they are asleep, already active or unwell.
5. Plan for two devices per person. Decide which source wins for each data type, deduplicate before scoring, and keep the source with every value.
6. Write claims you can defend. Act 3 features sit close to medical territory. Describe patterns and wellness, not diagnoses, unless the product has clearance.

What does the report not tell us?

It is a useful report, and a survey of 500 people has limits worth keeping in view:

• The method is thin on detail. a16z describes 500 randomly selected US consumers, but does not publish who ran the survey, when, the exact questions, or margins of error [1]. With 270 wearable users, each percentage among them carries roughly ±6 points of sampling error.
• The answers are self-reported. “Has your data ever prompted a change” depends on memory and on what people count as a change. Small, habitual adjustments are easy to forget.
• The two questions are not linked. The survey does not report whether the people who check most often are the ones who act, which is the most useful thing a product team could know.
• It is a consumer view. Employers, insurers and health plans, who fund much of digital health, barely appear.
• It is written by investors. a16z invests in consumer health, and the authors acknowledge that in some categories spending is moving faster than the evidence [1].

None of that undermines the headline. A majority of engaged, daily wearable users saying the data has never moved them is consistent with what the trials show at each step in between.

The State of Longevity describes a consumer who is engaged, curious and willing to wear more sensors. Its survey describes the same consumer looking at their data every day and, more often than not, doing nothing with it. The companies that close that gap will be the ones that treat data as the start of a decision, not the end of a dashboard.