From messy app data to numbers everyone agrees on

The situation

A consumer app collected a lot of information about how people used it. Product managers, engineers, sales and marketing all wanted answers from that data, each for their own reasons.

The problem

The information was stored the way the app needed it, not the way an analyst needs it: bundled together in large, nested records that were hard to read. Every report started with someone untangling the records by hand, and no two people did it quite the same way. The result was familiar. Two reports about the same thing would show different numbers, and meetings turned into arguments about which one was right.

What we did

The team rebuilt the data into a clean, standard layout of the kind used in well-run analytics set-ups, where customers, activity and products each have their own tidy table and link to each other in a predictable way. Anyone could now ask a question without first having to decode the raw records.

Just as important, the team sat down with product, engineering, sales and marketing to agree what each key number meant, and wrote it down. A term like "active user", for example, now had one clear definition.

The outcome

Data discrepancies dropped by 90%, so reports about the same thing finally matched. Reporting became around 50% faster, and the cleaner structure made it possible to build richer views, such as treemaps showing how activity was spread across the product.

What it meant for the business

People stopped debating the numbers and started discussing what to do about them. Requests from different teams could also be prioritised against agreed goals, because everyone was working from the same facts.

Could this work for you?

If reports disagree, or every new question needs a special effort to answer, the fix is usually in how the data is organised and defined, not in another dashboard.