Research is cheapest the week before you build and most expensive the month after you launch.
A small amount, yes. Not a six-month study. Enough to find out whether people understand the thing and want it before you pay to build it properly.
How much research is enough?
Five conversations with the kind of people who would use it, and three people trying a rough version. If the same problem shows up twice, it is real. If nobody can explain why they would use it, stop and rethink.
What should I ask in those conversations?
About what they do now, not what they would like. "Walk me through the last time you did this" gets the truth. "Would you use an app that did this?" gets politeness.
What if I have no time for research?
Then do the smallest version: show a sketch or a clickable prototype to three people this week. An hour each. It is still faster than rebuilding after launch.
When can I skip it?
When you are rebuilding something that already works and the customers are the same, or when the change is small enough to undo next week. Then launch it and watch the numbers.
Answered by Jared Lukes, Minneapolis. Twenty-five years in brand, UX, and technology, currently holding fractional AI and technology seats at three companies. About Jared.