"In analytics, the term fishing expedition refers to a project that was never framed correctly to begin with and involves trolling the data for unexpected correlations."
There is an excellent article in the MIT Sloan Management Review, Framing Data Science Problems the Right Way From the Start. This article is about data science, but it applies to any discipline. Most of the content will seem familiar and think of it as a refresher course on things you already know but may not be practicing.
Defining the problem needs to be step one. You can use the following checklist:
- Get the right people involved.
- Recognize that rigorously defining the problem is hard work.
- Don’t confuse the problem and its proposed solution.
- Understand the distinction between a proximate problem and a deeper root cause.
Do not move past the problem definition until it meets the following criteria:
- It does no harm.
- It considers necessary constraints.
- It has an accountability matrix (or its equivalent).
- It receives buy-in from stakeholders.
I've seen this work well with a design thinking approach and a simple design brief to get started.