I see this opinion a lot, but I disagree. Building things is precisely *how* you figure out what the conceptual data model is. It is an inescapable part of the process. https://twitter.com/kimgoodwin/status/1249802017301598208

If you like, you can start with a theory of what the conceptual data model is - a theory can be helpful in getting off the ground. But be ready to let it go almost as soon as you start building.

Be ready to adjust: to add and remove elements, to split concepts you thought were the same, and to combine ones you thought were different.

In fact, conceptual changes are probably the ONLY thing worth planning for ahead of time.

If you must plan: plan to make those sorts of changes easy.