Phase 1

Business Understanding

Turn a vague business problem into a precise question.

What this phase actually is

Business understanding is where a data science project earns its right to exist. The team turns a broad concern into a decision that can be supported with evidence.

That means naming the target outcome, the people who will use the result, the action they can take, and the cost of being wrong. Without that frame, even a technically elegant model can become a distraction.

The useful output of this phase is a precise question: who or what are we predicting, by when, using what data available at that time, and for which business action?

How this looks at Bertelsmann

Try it

Question Translator

Precise data question Which subscribers are likely to churn in the next 14 days, so the team can send retention offer? Target: churn risk. Action: Send retention offer. Window: Next 14 days.

Pitfalls

Starting with a model idea before agreeing on the business decision.
Optimizing a metric that nobody will use in the real workflow.
Treating "more data" as a substitute for a sharper question.