When it comes to fitness trackers, few people have more expertise than Tim Rosa. As the former CMO of Fitbit, he researched the reliability of the data they provide constantly, both individually and in larger studies.
Given his expertise, Rosa felt like the right person to approach to learn more about the reliability of fitness tracker data. He’s also deeply involved in sleep research in his current role as CEO of Somnee, so we asked him a wide range of questions about fitness tracker reliability, how sleep-related wearables are becoming part of the information equation, and the evolving role of AI in this technology. The following is everything that Rosa shared with us.
The lowdown on fitness trackers
The Manual (TM): Which fitness tracker features are the most and the least effective?
Tim Rosa (TR): To summarize, it varies across the board. The effectiveness of fitness tracker features depends on sensor location, the software models, and how well they align with your goals. Heart rate and activity tracking tend to be beneficial for those looking to monitor overall physical health trends, as they’re well validated and provide reliable data over time.
Beyond that, if improving sleep is your priority, then understanding your sleep patterns, your sleep stage trends, and your sleep score can be useful for identifying patterns and assessing the impact of lifestyle choices. Right now we track sleep in trends rather than stages, because you can’t accurately track sleep stages downstream from the head.
For some people, though, focusing too much on fitness and sleep scores can have the opposite effect, causing stress over bad readings and ultimately worse performance throughout the day. The same can happen with recovery scores, which are typically calculated through metrics like HRV, resting heart rate, and respiratory rate.
There’s another interesting part of this when it comes to athletic training, although it’s typically highly subjective. These scores can introduce a mental block for athletes who need to perform, regardless of whether the tracker says they should rest or not.
Meanwhile, we also have nascent tracker metrics — like AGEs (i.e., fitness age), which aim to give wearers info about biological aging. These are a new and somewhat interesting idea, but they may not provide particularly actionable data to the typical user.
Basic features like accurate GPS tracking, high battery life, and recovery metrics, for example, might be most important for higher-performing athletes and those training for a marathon.
TM: What are the most important metrics fitness trackers provide?
TR: VO2 max has largely been recognized as one of the most important markers for full-body health, longevity, and reduction of mortality, regardless of the cause. It’s true that a clinical setting is the only way to get a precise reading of VO2 max, but many modern wearable brands, like Garmin, Apple, Fitbit, and WHOOP, offer an estimate that gives users a good idea of where they stand.
In addition to long-term health indicators, wearables can also alert users of more immediate medical emergencies, like an irregular heart rhythm that could be a sign of AFib, as well as features that could detect early signs of a seizure or stroke.
An important caveat, though, is that fitness trackers and wearables don’t diagnose health conditions on their own! Always see a medical professional if there are signs of an urgent or life-threatening health scare.
TM: What should we rely on them for? What’s out of bounds or less reliable?
TR: Fitness trackers are most effective when used to compare your own results over a period of time to pick up on trends, rather than viewing one isolated metric as the “end-all-and-be-all” due to the inconsistency of metrics.
I’ve even noticed this in my own experience if I’m wearing multiple devices simultaneously for research purposes, as they’ll almost never show the same results across the board. Instead, monitoring the same metric over time — for example, your resting heart rate or sleep score — can give you a good idea of whether you’re trending positively or negatively. Look at the big picture rather than these micro numbers that almost always have some discrepancies.
TM: What is the optimal body placement for fitness trackers to get reliable data? How important is this?
TR: Optimal placement of fitness wearables can vary depending on the type of movement and data you’re tracking. For example, while chest straps are considered the benchmark for accurately measuring and monitoring heart rate, most finger-based trackers are proven to be more accurate than wrist-based devices when measuring heart rate.
Personally speaking, though, it’s not a fun experience wearing ANY ring when lifting weights or doing CrossFit workouts, so there are always trade-offs based on your goals and what you like to do for exercise.
As I mentioned earlier, sleep devices worn on the head are the benchmark for accurately measuring and gathering data on one’s sleep, because it’s measured directly from the brain.
The takeaway is that it’s essential to know your health and fitness goals and what you like to do for
This will help you determine the best device, sensor accuracy, and AI effectiveness, the last of which is improving on almost a monthly basis. Then you can connect all of that to your ideal experience and come up with some kind of cost-benefit value. There is no perfect wearable right now, but there’s so much progress happening.
The skinny on sleep data
TM: How is sleep data being integrated into trackers as a fitness marker?
TR: Sleep is sometimes considered a more nontraditional metric compared to heart rate, VO2 max, and other basic metrics, but the correlation between sleep quality, performance, and recovery in exercise and athletics is critical and validated by years of research.
Sleep is now considered a pillar of health and a key factor in achieving individual fitness goals, so we’ll continue seeing more interesting sleep data and performance and recovery correlation, along with more meaningful and actionable insights in wearables.
The question right now is about accuracy. At Somnee, one of our biggest achievements was having the NBA select our product as the only wearable or sleep technology to qualify for their Launchpad program. They’re using our trackers on players, coaches, and referees to create clinical-level sleep data as part of their performance platform, which shows how sleep is being taken seriously as a critical metric for athletic performance.
This data is constantly improving as technology continues to develop. And one thing we’ve learned is that wearable devices worn on the wrist/finger don’t provide entirely accurate sleep stage readings. Some perform better than others, so users should consider wearing these devices to understand “sleep trends” over time. They can be used along with sleep journaling to understand and improve basic sleep hygiene, but ultimately, the most accurate data will be provided by head-based devices that use lab-grade EEG sensors worn on the head.
A tech and AI assessment
TM: How accessible and straightforward is the tech, and how can this be improved?
TR: There is a fitness tracker/wearable out there to suit the goals of almost anyone today. Depending on your
At the highest level, there’s technology that provides highly intuitive and more complex solutions that are really suited to the needs of professional athletes.
Overall, the increase in consumer demand for more accurate data and health/fitness tracking is driving all this, along with the advancements in AI. More options are becoming available, along with competitive pricing that will make this technology more accessible to an increasing number of consumers.
TM: There’s a lot of interest in AI across the board, and in this area in particular. How would you assess its impact to date?
TR: Artificial intelligence is the game changer we’ve been waiting for with wearables. Although it’s still in its early stages, AI will fundamentally transform and advance wearable technology to what we want and expect.
In this area, I would expect AI to become more integrated into daily life. Wearables will become less visible, with less friction, and wearable AI will ultimately become the foundational tool that you can use as your personal doctor and real-time trainer, coach, and therapist.
There are also exciting new AI features coming online in wearables that measure things like holistic biometric data, lifestyle decisions, and so on. They’ll eventually be able to generate personalized recommendations, address questions, and provide responses to health and fitness goals, then come up with real-time plans to achieve short-term and long-term goals and outcomes.
Beyond that, the ultimate achievement with AI technology would be to produce a wearable that doesn’t just report accurate sleep data and insights, but also uses this data with ML/AI models to neuro-modulate the brain for better sleep quality, faster sleep onset, longer duration, and deeper restorative sleep. The whole phenomenon isn’t quite as existentially extreme as something like a William Gibson novel, but the deeper integration of AI and wearables is exciting and incredibly positive.
