Should You Build It? How to Validate a Side Project Before Sinking Months In
AI made building cheap, which quietly made validation MORE important, not less — because now you can build the wrong thing four times as fast.
There's a new failure mode among AI-era builders: shipping five polished products a year that nobody wanted. When building took six months, the pain of wasted effort forced people to check demand first. Now that a prototype takes a weekend, it's easier to "validate" by building — and end up with a portfolio of ghost towns. Cheap building didn't make validation obsolete. It made it the whole game.
What validation actually means
Validation is not your mates saying "cool idea." It's not upvotes, waitlist signups, or your own mounting excitement. Validation is evidence that specific people will exchange something scarce — money, time, or reputational risk — for the thing. Everything else is encouragement, and encouragement is free precisely because it's worthless.
The ladder of proof (climb until something breaks)
- The problem interview (cost: a few hours). Find five people in your target group and ask about the problem — never pitching the solution. You're listening for two things: do they describe the pain unprompted, and what are they doing about it today? "That's annoying but I live with it" kills more good-sounding ideas than any competitor. If they've duct-taped a spreadsheet solution — excellent. People who've built workarounds are people who'll pay.
- The landing page test (cost: a weekend). A one-page site describing the product as if it exists, with a price and a call to action (see our landing page guide). Send 100–200 of the right people to it via communities or small ads. A few percent clicking "buy" on a priced page is signal; a big waitlist from a page with no price is mostly noise — free email addresses measure curiosity, not demand.
- The pre-sale (cost: a hard conversation). The gold standard. "It'll be ready in six weeks — founding customers get it for $X, locked forever." Even three strangers pre-paying is more evidence than three hundred followers. If you can't bring yourself to ask for money before building, notice that — the discomfort usually means you don't believe the value case yet, and that's the thing to fix first.
- The concierge version (cost: some evenings). Deliver the outcome manually before automating it — the "newsletter picking stocks by hand" stage. Slow, unscalable, and the single richest source of product insight available, because you personally feel every step the software will need to do.
Signals that predict revenue (and the impostors)
- Real: pre-orders; people chasing you for launch dates; users of your manual version complaining when it's slow (they've integrated it into their lives); strangers — not friends — converting on a priced page.
- Impostors: launch-day upvotes (novelty traffic, gone by Thursday); waitlist size; compliments from people who've never had the problem; and the most seductive one — your own certainty. Founders are the only demographic that's 100% convinced pre-launch, every time.
Kill criteria: decide before you're attached
Write down, in advance: what evidence, by what date, keeps this project alive? ("Five pre-sales by March 1st" — whatever fits.) Side projects die properly or become zombies that eat every weekend for two years. A pre-committed kill line turns "giving up" into "following the plan," which is the difference between quitting a project and quitting on yourself.
And when the signal is real — move
Validation has a shelf life. If strangers are pre-paying and interviewees are chasing you, that's your green light to build fast and scrappy — this is exactly where the weekend-prototype tools earn their keep, followed by the launch checklist. The founders who win this era aren't the best builders or the best validators. They're the ones who do the two in the right order.
Build cheap, validate cheaper — and never confuse applause with a purchase order.