Challenge: An EdTech project faced a challenge: increasing conversion rates from free courses like 'Introduction to Programming' to paid programs. Despite effective student attraction and discounts, conversion stalled at about 7%. The goal was to boost these rates through strategic adjustments in course workload.
Solution: I analyzed the performance of 25,000 students in introductory course tasks, considering factors like completion accuracy, attempts taken, and hint usage. This led to the creation of a 'Learning Performance' metric for each student. By categorizing students into four groups — Frustrating, Challenging, Optimal, and Boring — based on this metric, I identified that students facing occasional challenges (the 'Optimal' group) had significantly higher conversion rates. In response, I recommended restructuring the challenging tasks, which were a common hurdle for many, into more manageable stages, aiming to increase overall task completion and engagement.
Impact: After adjusting tasks, most students shifted to the 'Optimal' group, leading to an increase in conversion rates from 7% to 18%. This approach balanced learning experiences, boosting student engagement and business outcomes.

Challenge: An EdTech project faced a 54% student retention rate beyond the first two weeks. Concerns about course design and teaching methods arose. We aimed to uncover factors behind this dropout rate, focusing on the balance between theory and practice.
Solution: To tackle this challenge, I developed the 'Cognitive Feasibility' metric, a tool designed to measure and assess the difficulty levels of both the theoretical content and practice tasks. This analysis revealed that a considerable number of drop-offs were linked to excessively challenging lessons. In response, we recalibrated these lessons to an optimal difficulty level. Moreover, we introduced a gradual increase in complexity, starting with more straightforward content in the initial weeks and progressively becoming more challenging, ensuring it remained within a manageable range for students.
Impact: Post-implementation, the first two weeks' retention rate improved to 87%. The overall completion rate also rose to 64%. These changes reduced student complaints and created a more effective learning environment.

Challenge: In an EdTech project, students lacked visible progress, causing dissatisfaction. They couldn't gauge their skill improvement, leading to uncertainty about their competency and the effort needed for mastery.
Solution: A 'Mastery' metric was developed, informed by students’ performance in the automated practice (considering accuracy, number of attempts before success, and use of supplementary resources) and project execution evaluated by expert reviewers. This metric hinges on two key indicators: the percentage of completed material for a specific skill, adjusted for the quality of learning performance. The Mastery metric thus not only quantifies progress but also assesses the depth of understanding, highlighting areas needing further practice. Consequently, this metric aids in tailoring the learning experience for comprehensive skill acquisition.
Impact: The Mastery metric significantly increased student trust, with a 23% rise in positive course evaluations. The Net Promoter Score also improved by 15%, reflecting enhanced student satisfaction.

Challenge: An EdTech project struggled to identify areas for content refinement. Relying on student feedback proved insufficient, as many students couldn't provide specific insights.
Solution: To address this, I introduced a 'Cognitive Feasibility' metric, grounded in how students interacted with practice tasks and test assignments. This metric, encompassing data on attempt frequency, time spent, and use of educational aids, offered an objective measure of each task's difficulty. By applying this metric across various levels — from individual tasks to overall topics — we could accurately identify and target challenging areas. This approach shifted the focus from subjective feedback to data-driven insights, enabling precise content refinement.
Impact: This metric guided monthly or quarterly planning of content improvements. It allowed for a balanced content team and substantially reduced expenses, improving educational content effectively.

Challenge: As student numbers grew in an EdTech project, scaling the support team linearly became inefficient. Our task was to restructure support while maintaining quality and reducing costs.
Solution: Based on how students solved practice tasks and test assignments, I proposed the 'Delta metric'. This metric calculates the deviation in a specific student's performance from the average (or model) student's performance. Crucially, the Delta metric is calculated for each individual task, as well as at higher levels — lessons and topics. We identified Delta values where the absence of human support for certain students could lead to severe consequences, such as dropout. Conversely, we also determined Delta values where students could rely solely on automated support, such as hints and other mechanisms. Furthermore, we identified a range in which students required no support at all.
Impact: Applying the Delta metric, we reformatted support, intensifying it based on performance. Support costs were reduced by 23%, while student satisfaction slightly improved, maintaining high educational quality.