How fitness apps can incorporate personalization without writing code?

Summary

The gyms and health clubs all over the world remain closed due to COVID-19. To keep their users engaged, it is now crucial for them to provide personalized services via fitness apps. Caboom can help fitness app creators develop personalized user experiences through an end-to-end recommender system development expertise. Caboom can help reduce the cost of implementation of the recommender system while generating value for fitness platforms.

Introduction

Fitness apps and wearable trackers have become hugely popular as a way of staying fit recently. Not only do they allow you to track and log your own workouts, but many also have a built-in social network where you can follow fellow users and their workouts routines. These social connections can be a great way of measuring yourself against as well as learning from others. This is especially useful for someone just starting out in their fitness journey or looking to improve their workout performance by learning from other high performing users. 

Let’s for example consider a user Mike. He is a 30 year old software developer training to take part in his first half marathon (21 kms). Being a tech guy, he’s been using a fitness tracker band connected to a fitness app “RunTracker” to train for the run. So far, he has been able to do 5k runs consistently, but hasn’t met the next milestone of 10k.

Fig 1: Data tracked by the RunTracker App


How can personalization help Mike reach his goal?

RunTracker wants to help Mike reach his goal of running a half marathon. But, he is struggling to reach the next milestone of a 10k run. There’s a risk that he might be demotivated to continue his training. However, there are other users of the app very similar to Mike who have gone through the same journey and have taken that next step to become successful half-marathon runners. 

The creators of RunTracker see this opportunity. They plan to add a personalization feature which allows users like Mike to discover and connect to other users who they can follow and learn from. They will also be able to request mentorship from these users if they accept that. 

Given the number and variety of users in their app, manually curating these recommendations for each user is not a practical option for RunTracker. This is a problem best solved using an AI-based solution. However, AI is new for the RunTracker team. This means that they would either need to hire a team of Data Engineers, Data Scientists and ML engineers and spend anywhere from 6 months to a year to implement this feature. Alternatively, they can come to the Caboom team and make the process a lot easier.

Fig 2: No AI. Users recommended randomly or based on simple rules


How Caboom is better than traditional AI development?

Caboom has expertise in developing AI recommendation solutions precisely like this. With Caboom they get the first version of the feature out in a matter of weeks not months. The main reason for this is our expertise and products designed to streamline this process from concept to production. 

The Caboom app is an easy to use no-code ML platform where users can easily learn about how a personalized recommendation system works, what kind of data is required to develop this kind of solution, and also try out creating a simple AI model with a sample of their own data. The users can simply take this model to their developers and implement the feature on their own. 

For cases like RunTracker who need a more customized solution the Caboom concierge service is a better fit.

Fig 3: Traditional AI-based Way. Required a huge development and integration effort

Runner Level Score

  1. Score similar to a golf handicap
  2. Convert the time, GPS (distance, speed) and performance data (heart rate, calories) of each run into a performance score
  3. Establish a target level to measure yourself (100 means an elite runner)
  4. Give a score based on recent runs that can be compared against an athlete one level above your current level
  5. The runner levels are differentiated in terms of warmup, stretches, posture and nutrition, that the users generally seek knowledge transfers on. 
  6. Breakdown performance in different terrain, altitude.

Suggested People to Follow

  1. Create candidate runners to show the user (Mike) in the app based on their history and current level.
  2. Filter the users based on Mike’s target score, location (similar routes) and other criteria he may have specified like similar age group etc.
  3. Show the best matches to Mike to follow and connect to.
Fig 4: The Caboom Way. Easy integration of Caboom Engine to simplify the whole process

Thus, with personalized recommendations like these, Mike can find other runners whom he can follow, compare himself with and also connect with for added advice and mentorship. This also brings in an element of gamification to the runs which is an added motivation.


Conclusion

Having personalization baked in RunTracker apps also improves it’s user experience and retention rates. It's also a valuable feature that can help in new user acquisition. Additionally, RunTracker can also create new revenue streams by providing more detailed analysis of a runner's performance based on the Runner Level model, and partner with running gear manufacturers and trainers. Thus, providing a win-win for both users and the app. 

This is just one way how Caboom can help healthcare and fitness apps make their app experiences better. They can easily get started by trying out our no-code, self service solution to see the feasibility of adding personalization to their products and then seamlessly upgrade to our concierge services for more advanced cases.


See instant AI recommendation results with caboom
Start with your data or a sample for instant results right away.
Request Access