Should you build an MVP or a landing page first?

The short answer

Landing page first, in almost every case. A landing page tests whether anyone wants the thing in days and for almost nothing; an MVP tests whether you can build it, which is rarely the real risk. Build the MVP only after strangers have signed up, preordered, or paid on the page.

This ordering question decides how expensive your first mistake is. Get it wrong and you spend three months building something nobody asked for; get it right and the market tells you within two weeks whether the three months are worth spending.

Why the landing page usually wins

Each artifact answers a different question. The landing page answers 'does anyone want this?': a concrete promise, a price if you dare, and a button that asks strangers for something costly, like an email or a preorder. The MVP answers 'can I build this?', and for most software ideas in 2026 the honest answer is obviously yes, which makes building it first a way of avoiding the scarier question. Demand risk almost always exceeds build risk, so test demand first.

The exceptions

Build first when the product itself is the proof: a technical breakthrough where feasibility is genuinely uncertain, or an experience so novel that describing it convinces nobody. Even then, the build should be the smallest demonstrable core, not a polished app. And a landing page with zero traffic proves nothing, so the page must come with a plan to put it in front of strangers; that whole motion is called a smoke test, and how to smoke test a startup idea walks through it, while what counts as real validation covers how to read the results.

In GRILLR, this ordering is built into the plan itself: validation tasks come before build tasks, the AI-run research phase sizes demand before you write code, and you cannot mark the landing page task done by feeling optimistic about it. You submit the live link, and it either passes or it does not.