How does Tutlo increase user engagement and motivate them to learn with Machine Learning models?

Tutlo is a dynamically developing educational platform specializing in learning English online. The company offers modern digital solutions, connecting teachers with students in a flexible and convenient way. To ensure the highest quality of service and operational efficiency, Tutlo constantly invests in innovative technologies.

Thanks to cooperation with Alterdata, Tutlo has gained the opportunity to better understand the behavior of its users and more effectively increase their engagement and motivation to learn by using Machine Learning models in the BigQueryML environment

Project objective:

Tutlo needed a deeper understanding of student behavior in a flexible learning model and tools to predict their engagement and motivation to learn. The key was to use machine learning models to identify factors influencing user activity, create measurable KPIs, and implement a scalable analytical solution that would enable faster business decisions and more effectively increase the satisfaction of both students and teachers.

Scope of work

1. ML consulting and modeling:
   • Our Alterdata team conducted workshops with the client to thoroughly understand how, among others: flexible teacher selection in real time affects students' engagement and motivation to learn.
   • We have developed a custom solution, tailored to Tutlo's unique business model, which allows for even better technology matching to the specific needs of data integration and behavioral analysis using BigQueryML
     
2. Creating KPIs:
   • Working closely with Tutlo, we developed over 80 variables and performed advanced user segmentation to better understand their needs and how they use the platform. Thanks to this, it was possible to precisely adjust activities that increase the comfort of learning, improve user experience and make using the platform even more intuitive and satisfying.
     
3. Application of technology:
   • We implemented the XGBoost model in the BigQueryML environment, which enabled full integration of the Machine Learning process with the data storage location. This allowed the model to pull data directly from BigQuery and return results directly to the same platform, eliminating the need for transport between systems.
   • This approach significantly simplified model management, increased its scalability and eliminated delays resulting from data migration, which resulted in more effective and smooth operation of the entire solution.

Implementation stages

1. Analyzing and defining goals
   • We determined with the team key variables influencing user engagement.
   • We conducted joint workshops that included both the analysis of the client's business model and its consequences. The result was the development of model use scenarios that took into account various business needs.
     
2. ML model design
   • Our team has developed a model that precisely predicts the probability of specific events.
   • We performed econometric customer segmentation, adapting predictions to different user groups.
     
3. Technology implementation
   • We implemented the solution in BigQueryML, which enabled:
     • time savings during model implementation and maintenance,
     • scalability and minimization of operating costs.

Results

Key results:

• Model accuracy: Achieving 80% accuracy in predictions.
• Technological simplification: Implementing the model in BigQueryML allowed for easy management and no need to transfer data to external environments.
• Growth business awareness: The client understood which factors have a key impact on user engagement, which allowed the development of new methods of motivating and activating users. 

Business value:

• Long-term benefits: Increasing users and teachers satisfaction
• Process optimization: Reduce decision-making time with data-driven forecasts.

Client feedback

The solution provided by the Alterdata team was extremely precisely tailored to our needs. Thanks to BigQueryML, we were able to develop the model ourselves and better understand the key factors influencing the success of our platform.

The solution provided by the Alterdata team was extremely precisely tailored to our needs. Thanks to BigQueryML, we were able to develop the model ourselves and better understand the key factors influencing the success of our platform.

Jacek Niedzielski,

Data Analyst at Tutlo

Summary

The project implemented for Tutlo shows how experience, knowledge, and innovative technologies can build competitive advantage in practice. The BigQueryML-based solution was a breakthrough in managing student engagement and improving relationships with teachers.

Our ability to create personalized ML solutions has confirmed our position as a leader in delivering valuable data-driven solutions.

If your company wants to better understand user behavior and use data to increase engagement and operational efficiency, contact us and let's talk about your project.