A French healthtech scaleup turned LTV into a daily decision tool. From board-slide metric to revenue driver.

A leading French healthtech scaleup serves millions of users across a freemium-to-paid funnel. In a business that mixes broad acquisition with long retention, knowing what each customer is actually worth, not on average, but individually, is the difference between scaling efficiently and burning cash on the wrong cohorts.

The company had the data, the engineering maturity, and a strong analytics team. What it didn't have was a way to turn LTV from a quarterly board-slide metric into an operational tool used every day.

The scaleup's leadership was making thousand-decision-per-week tradeoffs, acquisition channels, retention spend, segment prioritization, lifecycle investments, with a single LTV number computed as a historical average. The math looked clean. The reality was that some users were worth ten times the average and others a tenth, and the company was paying the same to acquire all of them.

Building a real predictive LTV system meant either staffing a data science squad with strong product-ML chops, or running a long external project. Both options would push back the ROI by quarters. And in a scaleup that ships weekly, quarters are a long time.

Matr gave the scaleup a way to put a predictive LTV model in production without growing the team or starting a parallel data program. The analytics team owned the work, the modeling complexity moved to Matr, and the business teams got to use the output where decisions actually happen.

What sold the team was the combination: a model trained on their warehouse data, an output usable by non-technical functions in plain language, and an explainability layer so the CRO could trust the number when steering revenue, not just look at it.

The scaleup's data team connected Matr to their warehouse and built the first predictive LTV model in days, with no dedicated data engineering effort. Each user gets an individual predicted value, refreshed continuously as their behavior on the platform evolves.

From there, the data team operates Matr the way they'd operate a data science team that never sleeps: defining the LTV objective, validating the modeling approach, monitoring drift, retraining when behavior shifts. The model stays accurate over time, with no parallel infrastructure to maintain.

This is what made predictive LTV useful rather than just accurate. The marketing, growth, and revenue teams don't open Matr or interrogate model internals. They consume the output where they already work.

Marketing uses predicted LTV to bid harder on high-value lookalikes and rein in spend on cohorts the model flags as low-value. Growth prioritizes retention investments based on the predicted upside per user, not on a flat assumption. The CRO uses LTV as a steering metric across channels, segments, and quarters, with the model's reasoning available whenever a decision needs to be defended.

The result isn't a one-off LTV calculation. It's a working capability: a data team running predictive LTV in production, and revenue-side functions making daily decisions on it.

• +15% acquisition ROI by reallocating spend toward high-LTV segments
• 3 days to build and deploy the first model in production
• 5 million LTV predictions refreshed monthly across the user base
• Used daily by marketing, growth, and revenue leadership, with LTV operating as a steering metric, not a quarterly report

Today, the scaleup runs predictive LTV alongside other Matr models on the same platform, with the same data and the same explainability lay

‍