You will own and develop the LTV prediction algorithm by prototyping improvements in SQL and validating forecasts against actual data. Additionally, you will conduct end-to-end analytical research to provide actionable recommendations for Product, UA, and Finance teams.
Senior Data Analyst (LTV / Predictive Analytics)
StellarTech is a B2C mobile publisher in EdTech: subscription apps, in-app monetization and performance marketing at scale. LTV prediction is how we decide where to invest, and the forecasting pipeline is already live. We are opening this role to have one senior analyst own its accuracy and its evolution. This is research and predictive analytics, not reporting: everything that can be standardised is already automated by Data Engineering, so the work is hypotheses, factor analysis and improvements to the prediction algorithm itself. You report to the Head of the Operational Department and work directly with Product, UA, Finance and Data Engineering.
What you will do
Own and develop the LTV prediction algorithm: find the parameters and factors that increase forecast accuracy, prototype improvements in SQL, measure the gain, then hand the logic over to Data Engineering.
Validate forecasts on closed cohorts, monitor forecast quality over time and explain where exactly a model deviates from actual LTV.
Research what moves LTV: product features, segments, purchase sources (for example, purchases coming through CRM vs other sources), and adapt the prediction logic to them.
Run analytical research end to end, from a vague question to a documented recommendation for Product, UA and Finance: where payback improves, where to concentrate media buying, what changes unit economics at company level.
Automate analytical routines and model quality monitoring together with Data Engineering.
Requirements
Critical
4+ years in Data Analyst / Senior Data Analyst / Product Analyst roles with a product analytics focus: measuring the impact of factors on business metrics, not marketing attribution.
Strong SQL: complex queries, window functions, optimisation. Here SQL is the main instrument for prototyping and testing the algorithm.
Solid mathematical statistics and understanding of how ML models work: you can build and validate a forecast, judge model quality (MAE, RMSE, MAPE, R2), run factor analysis and explain where a model broke.
Product background with in-app monetization and its economics: LTV, cohorts, retention, ROI, CAC, ARPU, payback. Subscriptions are ideal; other in-app B2C models also work. Ad-only monetization is a different pattern and is not a fit.
A proven end-to-end pattern: you take a vague question, form a hypothesis, dig into it and bring it to a recommendation. There is no micromanagement here and almost no standard tasks.
Ability to back conclusions with data and to communicate them to both technical and business stakeholders.
Non-critical
Python (pandas, numpy, scikit-learn) for analysis and modelling: welcome, but not a blocker if the statistical and analytical thinking is there.
A/B testing and statistical hypothesis testing.
BI tools (Tableau, Looker, Power BI).
Cloud DWH and large data volumes (Athena, BigQuery, Snowflake, Redshift or similar).
Will be a plus
Feature engineering and model interpretation (SHAP, feature importance).
Mobile products, gaming or subscription businesses.
“I was the first applicant for a remote marketing position that got listed on the company website the same day I applied. Had an interview within 48 hours!”
— Sarah J.
·
Marketing Manager
★★★★★
Verified
