Google Fit tends to undercount your steps, often by a significant margin. A 2025 validation study found that Google Fit undercounted steps by 20% to 60% depending on walking speed and how long the activity lasted. That puts it well outside the accuracy thresholds researchers use to evaluate fitness trackers, which are typically within 3% error in controlled settings and 10% in everyday life.
How Much Google Fit Undercounts
The most detailed look at Google Fit’s step accuracy comes from a 2025 study published on medRxiv that compared the app’s counts against manually counted steps. The results were consistent: Google Fit underestimated steps across every test condition. At slower walking speeds, the undercounting was worst, reaching 60% in some trials. At faster, more rhythmic paces, the gap narrowed but still hovered around 20%.
The problem wasn’t random. A statistical analysis showed a consistent negative bias, meaning Google Fit didn’t just miss steps occasionally. It systematically reported fewer steps than actually occurred. On top of that, the trial-to-trial variability was substantial, so the size of the error wasn’t predictable from one walk to the next. You might be undercounted by 25% on a morning walk and 45% on an afternoon stroll, even if you walked the same route at a similar pace.
For context, researchers generally consider a step counter “accurate” if it falls within 3% of the true count in a lab setting or within 10% during normal daily activity. Google Fit’s 20% to 60% undercounting falls far outside both thresholds.
How Google Fit Detects Steps
Google Fit relies on your phone’s built-in accelerometer and gyroscope to detect steps. The accelerometer measures the intensity of motion across three axes, and the app calculates the overall magnitude of that movement to identify the repeating pattern of a footstep. This approach works regardless of how you hold or orient your phone, since it’s looking at the total force of each motion rather than movement in a single direction.
GPS plays almost no role in step detection. GPS signals degrade indoors and in urban environments with tall buildings, so the app leans on those motion sensors instead. GPS data may help Google Fit estimate distance for outdoor activities, but the step count itself comes from the accelerometer.
This sensor-based approach explains why accuracy drops at slower speeds. When you walk slowly, the acceleration produced by each step is smaller and less distinct. The algorithm has a harder time separating genuine steps from background noise, like shifting your weight or fidgeting. At a brisk, steady pace, each footfall creates a sharper, more recognizable signal.
Where Your Phone Lives Matters
Because Google Fit depends on detecting motion through your phone, where you carry it has a direct effect on accuracy. A phone in a front pants pocket moves in sync with your leg and produces a clean, rhythmic signal with each step. A phone in a handbag, on a desk, or in a loose jacket pocket picks up weaker or more erratic motion, leading to more missed steps.
The reverse problem also exists. Certain repetitive motions can register as false steps. Riding in a car on a bumpy road, rocking in a chair, or even vigorous hand gestures while your phone is in your hand can generate acceleration patterns that look enough like walking to fool the algorithm. Google Fit does attempt to filter these out. According to Google’s developer documentation, the app detects activities like cycling, driving, and public transit and removes step counts that were logged during those periods. But the filtering isn’t perfect, and short bursts of vibration or movement may still slip through.
How Multiple Devices Are Handled
If you use both a phone and a Wear OS smartwatch with Google Fit, the platform merges step data from both sources rather than simply adding them together. Watch data takes priority over phone data, which helps avoid double-counting steps that both devices recorded simultaneously. The merged result is meant to reflect the most complete picture of your activity.
In practice, you may notice temporary mismatches between your watch and phone displays. These are usually syncing delays. Google’s backend holds the most current merged data, and both devices will eventually show the same number once they sync. If you consistently see different totals, it’s worth checking that both devices have an active connection and that Google Fit has permission to access motion data on each one.
What This Means for Your Daily Count
If you’re using Google Fit to track whether you hit 10,000 steps in a day, the undercounting means your real total is likely higher than what the app shows. A displayed count of 7,000 steps could represent anywhere from 8,500 to over 11,000 actual steps, depending on your walking speed and how you carried your phone. That’s a wide range, and it makes Google Fit unreliable for precise step goals.
Google Fit is still useful for tracking relative trends over time. If your step count goes up week over week, you’re genuinely moving more, even if the absolute number is off. The bias is consistent enough that comparisons within the app hold up reasonably well. But if you need accurate counts for a rehabilitation program, a research study, or a step-based challenge with friends using different devices, Google Fit’s numbers will run lower than most dedicated pedometers and fitness bands, which typically land closer to that 3% to 10% accuracy range.
A few things can help you get the best results from Google Fit. Keep your phone in a pants pocket close to your body. Walk at a steady, moderate-to-brisk pace rather than shuffling. And treat the number as a rough estimate rather than a precise measurement.