In the fitness app landscape, accurately detecting a user’s activity level is essential – not just for tracking performance, but also for ensuring fairness and boosting engagement. One of the key features that are analyzing today is speed detection, which classifies user activity into walking, medium run, or fast run.
As a Business Analyst, I’d like to walk through how this feature works, what problem it solves, and where its limitations lie – and how we can make it even better.
The goal of the feature
This feature helps the app:
- Track and log workouts accurately.
- Provide appropriate rewards (e.g., points redeemable for cash, digital coins,…).
- Prevent cheating or misuse in gamified fitness challenges.
- Offer more personalized insights and coaching.
How the speed detection works
The system uses a combination of:
- GPS speed: Measures how fast the user is moving.
- Phone vibration sensors (accelerometer): To detect foot impact or motion when the phone is held or in the pocket.
Based on GPS speed and phone vibration data, the system detects activity as follows:
- Walking: is detected when the speed ranges between 1 – 6 km/h with rhythmic vibration.
- Medium running: is detected when the speed ranges between 6 – 10 km/h with rhythmic vibration.
- Fast running: is detected when the speed ranges between 10 – 18 km/h with rhythmic vibration.
- If the system records a very high speed without noticeable vibration, it will flag this as potential motorbike cheating.
These thresholds help the system categorize movement more accurately and support features like progress tracking, coaching tips, or in-app challenges.
Problem: Smart cheating
Some users try to game the system:
- They ride a motorbike to quickly rack up distance or points.
- To mimic “vibration,” they shake the phone by hand.
This exposes a limitation: GPS + vibration alone may not be enough to guarantee accurate detection.
How the feature can be enhanced for better accuracy
To improve accuracy and reduce false positives/ negatives, we can enhance the feature by:
- Syncing with Smartwatch or Fitness Band:
- Track heart rate to validate physical effort.
- For example: If speed is high but heart rate remains low, it’s likely cheating.
- Motion Pattern Analysis:
- Use AI to recognize consistent footstep patterns vs. erratic hand shaking.
- Contextual Awareness:
- Detect route types (e.g., sidewalk vs. main road) to further validate activity.
What makes this feature valuable for the business
Improving this feature adds value across multiple dimensions:
- Accuracy → Better fitness tracking = higher user satisfaction.
- Trust → Fair competition = better community engagement.
- Retention → When users feel achievements are meaningful, they’re more likely to stay.
Conclusion
As a Business Analyst, I believe it’s essential to balance user experience, technical feasibility, and business goals. The speed detection feature is a great example of how we can build on a smart foundation and continuously enhance it through data, sensor integration, and insights into user behavior.