When success is vague, turn the ambiguity into a set of possible performance dimensions such as profitability, growth, customer outcomes, market position, or operations.

Tell the interviewer that without a single target, you would first diagnose which dimensions are underperforming and which matter most to the client.

That gives you a structure that can narrow toward a recommendation as evidence appears. If useful, this guide on case structure vs. case framework gives more detail.

I turn 'performance' into possible metrics first, then diagnose.