New sellers usually ask which category sells best. It is the wrong question, and the right one is more useful.
Solving one problem completely beats solving five partly
The items that keep selling tend to do one thing thoroughly. An invoicing script that handles tax, part payments, reminders, PDFs and the awkward edge cases sells better than a "business suite" that does eight things at surface level.
Buyers arrive with a specific problem. They are searching for that problem, not browsing for a suite.
Boring beats exciting
Authentication, billing, admin panels, imports, reporting, notifications. Nobody is excited about them, everybody needs them, and nobody wants to build them again. The unglamorous middle of an application is where sustained sales come from.
Being current matters more than being clever
An item that keeps up with the versions its buyers are on holds its sales for years. One that was excellent for a framework two majors behind stops selling regardless of how good it is. Regular small updates are the single strongest signal that an item is alive.
The author is part of the product
Look at any item that has sold consistently for years and you will find an author who answers comments. Buyers can see the whole support history on the page before they decide, and they read it. This is the part that cannot be copied by a competitor.
Niches are not too small
"A booking system" competes with everything. "A booking system for driving instructors, handling test dates and block bookings" competes with almost nothing, and the people who need it will pay properly for it, because the alternative is a custom build.
The compounding part
Items rarely take off at launch. They accumulate: a few sales, a few comments answered, an update, a review, better placement, more sales. Sellers who quit at eight weeks quit right before the curve. The ones still there after a year are usually the ones still shipping small updates.
