By Pranav Bhaven Savla
A few years ago, I started doing something small and slightly embarrassing. On mornings when I woke up feeling tired, I would check my phone to confirm it. On mornings when I woke up feeling fine, I would also check my phone, just to be sure.
At first, this felt harmless. Sensible, even. If technology could tell me something useful about my sleep, why not listen? But over time, I noticed a shift. The question quietly changed from “How do I feel?” to “What does the data say I should feel?” That question now shapes much more than sleep.
Over the last decade, performance has been quietly rewritten in numerical terms. Sleep is scored. Movement is counted. Work is tracked. Focus is timed. Progress is graphed. These measurements don’t arrive as commands. They arrive as suggestions, summaries, gentle nudges. And because they are grounded in real science, we tend to trust them.
To understand what’s happening, it helps to look closely at how these numbers are made.
Take sleep tracking, now one of the most common forms of personal performance measurement. Consumer wearables estimate sleep using motion sensors and heart-rate data. These signals are meaningful. Changes in heart rate variability and movement do correlate with different sleep stages across large populations. This is why the approach works at all.
But it’s also important to understand what these devices are not doing. In sleep science, the gold standard is poly-somnography, which measures brain activity using EEG, along with eye movement and muscle tone. Wearables do not measure the brain. They infer sleep indirectly.
When researchers at places such as Harvard Medical School, Stanford University, and the University of Oxford have compared popular consumer wearables with clinical sleep studies, they’ve found a consistent pattern. Total sleep time is often estimated reasonably well when averaged across groups. But accuracy drops sharply at the individual level, especially when distinguishing light sleep from brief awakenings. Night-to-night variability is high. The margin of error is large.
This uncertainty is well understood by sleep scientists. It rarely reaches users.
What reaches users is a single number. A score. A readiness indicator. Something that looks clean enough to trust. Around 2017, clinicians began describing a pattern that followed naturally from this design. Patients arrived at sleep clinics distressed by poor sleep data despite functioning well. They reported feeling anxious about their numbers, changing bedtime routines to chase better scores, and lying awake worrying about whether they were sleeping “correctly.” The clinicians called this orthosomnia: sleep disrupted by the pursuit of ideal sleep metrics.The data didn’t just describe sleep. It changed how sleep was experienced. Physical activity tracking shows how easily this shift spreads. The familiar goal of 10,000 steps per day feels scientific, almost medical. In reality, it began as a marketing choice for a Japanese pedometer in the 1960s. The number was memorable. That was its main qualification.
Only decades later did researchers test step counts at population scale. In 2019, a study published in JAMA Internal Medicine followed more than 16,000 older women for over four years. The researchers found that mortality risk decreased as daily step counts increased, but only up to about 7,500 steps per day. Beyond that point, the curve flattened. Walking more did not meaningfully reduce risk further. A 2020 study using data from the U.S. National Health and Nutrition Examination Survey found similar nonlinear patterns across broader age groups.
The science here is not complicated. Movement matters. More movement helps, but not indefinitely. Health depends on intensity, baseline fitness, age, and context. There is no single number that applies to everyone.
The interface cannot say that. Instead, it sets a target. And once a target exists, it becomes a standard. A long day spent lifting, standing, carrying, or working with your body can feel oddly unsatisfying if the counter stays low. A day spent pacing the house can feel successful if it crosses a threshold. The number starts to define the day.
This isn’t a design accident. It’s how feedback systems work. Economists call it Goodhart’s Law: when a measure becomes a target, it stops being a good measure. Engineers encounter it whenever a system rewards the wrong proxy. What’s new is that we are now applying this logic inward, to our own bodies and minds.
Work makes the consequences clearer. Most productivity tools measure what software can see: time active, tasks completed, response speed, interaction frequency. These signals are easy to collect and easy to compare. What they struggle to capture is thinking.
Cognitive science has shown for decades that deep reasoning and learning are uneven processes. Insight often follows periods of low visible output. Struggle, delay, and mental simulation are not inefficiencies; they are mechanisms. From the perspective of a dashboard, they look like nothing.
Research from groups such as Microsoft Research and Stanford’s Human-Computer Interaction lab shows that increased monitoring changes behavior. People respond faster, fragment work, and stay visibly active. The work becomes easier to track and harder to deepen. Output increases. Understanding does not necessarily follow.
If this sounds familiar, it’s because most of us live inside this tension. A day filled with emails and small tasks looks productive. A day spent thinking looks empty. Over time, you learn which days the system prefers. Health metrics add another pressure: the expectation of continual improvement. Many apps establish a personal baseline and encourage you to exceed it. This sounds reasonable until you compare it with biology. In exercise physiology, gains slow. In learning science, plateaus are normal. In sleep research, variation is inevitable.
Human systems adapt, stabilize, and recover. Software systems are optimized for engagement. Engagement thrives on change. Targets rise. Streaks reset. Stability starts to feel like failure.
None of this arrives as criticism. It arrives as numbers. And numbers feel neutral. Fair. Hard to argue with.
But metrics don’t see context. They don’t see care-giving, illness, emotional labor, stress, or recovery. Two identical scores can represent entirely different lives. What the system cannot measure disappears from evaluation.
This is the scientific mistake at the heart of the shift: confusing signals with meanings. A heart rate is a signal. Health is an interpretation. A step count is a signal. Fitness is an interpretation. A productivity metric is a signal. Value is an interpretation.
When numbers replace judgment instead of informing it, performance narrows.
I still look at my data. Most of us do. The problem isn’t measurement. It’s authority. Somewhere along the way, many of us stopped asking what a number actually captures and started treating it as a verdict. Science was never meant to work that way. Measurement is supposed to sharpen judgment, not outsource it.
The real skill we need now is not better tracking, but better reading. Knowing when a metric is useful, and when it’s just confident. Knowing when to take it seriously, and when to set it aside. Because a life well lived will always be messier than a dashboard. And performance, in the fullest sense, still lives partly in places that no sensor can reach.
(The writer is First Year Student, Plaksha University).
