How Start 2 Run Keeps Runners Injury-Free With AI-Powered Speed Recommendation
Running injuries are the primary reason people abandon their fitness journey, with studies showing around 50% stopping within their first half year following injury. Start 2 Run creators of Belgium and the Netherlands' (where it is called Hardlopen met Evy) largest running app, partnered with In The Pocket to develop an AI solution that would help prevent injuries by determining safe running speeds for beginners. By combining innovative machine learning and sport science, we created a system that predicts a person’s safe running speed, transforming how runners approach their training.
Noticing a high dropout rate due to running injuries, Start 2 Run wanted to help beginners run at safer speeds.
An AI-powered system that predicts safe running speeds and integrates seamlessly with the Start 2 Run app.
By analysing 7,500 running sessions and combining science with AI expertise, we developed an accurate speed prediction model.
The new app feature provides personalised speed guidance to help runners stay motivated and injury-free.
01 - CHALLENGE
Preventing Injuries For Start 2 Run App Users
Start 2 Run is Belgium and the Netherlands' leading provider of digital running guidance, which has helped millions of runners since 2013. The product’s owner, sports lab and corporate wellbeing provider Golazo Energy, combines sports science expertise with technology to create personalised training programs for runners of all levels.
Klaas De Waele, Product Manager at Golazo Energy, explains why running is one of the most popular sports. “Running is special because anyone can do it. You don't need an expensive gym membership or special facilities. All you need is a pair of good trainers and the motivation to begin. Plus, it's beneficial for both physical and mental wellbeing.
”However, there is a major issue when people start running without previous experience or professional guidance. “We know from the lab that 90% of new runners run too fast when they start. Many beginners try to run at 9-12 kilometres per hour, which is way too intense for someone just starting out. We needed a way to help them find the right pace – one that allows for progress whilst also preventing injuries," says Klaas.
Start 2 Run saw another critical issue with their users. “Research in Holland revealed a concerning trend: 30% of new runners quit within their first half year, mostly due to injuries or lack of motivation because they don’t see progress. Our app already covered a lot of the issues, but this showed us we needed to focus our approach even more for this audience," he shares.
While their sports lab can easily determine safe running speeds through physical testing, they needed a scalable solution that would transform collected data into personalised, real-time guidance for beginners.
02 - SOLUTION
Creating An Intelligent Speed Prediction System To Guide The Runners
“We collaborated on a project with Sportinnovator, a Dutch initiative funded by the Ministry of Health, Welfare, and Sport (VWS) and managed by ZonMw. We were selected as one of three grant recipients, enabling us to collaborate with a digital partner to help realise our ideas into an AI-powered solution, on top of many smaller features that we have brought together, focussing on injury prevention, through several app updates."
We interviewed different companies, but In The Pocket stood out. They showed a real interest in understanding the science behind running and how we work with our clients. This wasn't just about building a theoretical AI system, it was about creating something that would truly help our runners," says Klaas.
Our team has developed a machine learning solution that processes and analyses patterns in runners' speed and heart rate data to provide personalised pace recommendations."
One of the key aspects of not getting injured is finding the balance between training whilst maintaining reasonable load on your body and low fatigue to allow for recovery and ultimately building up stamina, fitness and strength," explains Klaas. That is why we trained the AI system to combine complex data analysis with user-friendly implementation, making professional-level guidance accessible to all users.
03 - APPROACH
Combining Sport Science with Machine Learning Innovation
Our approach combined AI with user-centred design to create a comprehensive solution. On the technical side, we developed a Long Short-Term Memory (LSTM) neural network that processes sequences of running data, using speed and heart rate measurements to make accurate predictions. This solution, deployed on their cloud backend, ensures reliable performance and scalability.
The technical solution is complemented by a carefully crafted user experience (UX) design that shows the information in an intuitive, user-friendly way through the Start 2 Run app. The designs include features like personalised running reports, hands-on speed recommendations, and engaging coach interactions, making the complex calculations accessible and actionable for beginners. Depending on analysis and feedback we aim to further improve the user experience for even better results.
04 - OUTCOME
Delivering Personalised Experience For Lasting Running Enthusiasts
The AI solution delivers impressive accuracy in predicting safe running speeds: almost 9 in 10 predictions fall within the margins we determined in lab tests. Combining multiple weighted results with further conditions and filtering, we reach a prediction accuracy that falls well within what we aimed for.
Launched at the start of 2025, this new app feature strengthens Start 2 Run's market position by providing a unique value proposition for new and existing users and improving retention through safer, more effective training. "This new feature will help runners find their ideal safe pace from the start. When you run at the right speed, you enjoy it more, make better progress, and are much less likely to get injured. We believe this will help more people stick with running and achieve their fitness goals," explains Klaas.
De Waele emphasises how success of this AI system is a result of great collaboration with our team, which built the foundation for future app improvements: "Working with In The Pocket was a true partnership. We combined our running expertise with their technical knowledge. Nina Zizakic and her team took time to understand our coaches' methods, analysed thousands of running sessions, and worked with us to create an AI system that could identify the optimal running speed for each person.
”The data-driven solution offers unique value to new users but also improves retention through safer training, while laying the foundation for future app features that will keep optimising the beginner running experience. For example, in the latest Start 2 Run updates, the Start 2 Run team put injury prevention first with Safe Speed AI and subjective fatigue monitoring for safer running. Inexperienced runners can now benefit from our 'Easy Buildup' plans and smart triage system, while all users get access to pain tracking with targeted exercise recommendations.
Through this innovative collaboration, Start 2 Run and In The Pocket have not only revolutionised how beginners approach running but also set a new standard for AI-powered fitness guidance, proving that when technology and sports science work together, everyone can achieve their goals safely and confidently.
