Most idea validation is theater. A maker has an idea, asks five friends if they'd use it, hears "that sounds cool," and builds it. Six months later the product launches to silence. The "validation" was indistinguishable from the absence of validation.
Real validation is narrower, faster, and more uncomfortable. It produces signal because it asks the question that matters: will someone trade something they care about (money, time, an email address) for the idea before it exists? The methods below all answer that question without writing the product.
The four methods that actually work
Four cheap, fast validation methods reliably produce signal. Each costs between $0 and $200 and runs in one to three weeks.
The pre-built landing page. A single URL describing the product as if it exists, with an email capture and one CTA. The signal: how many of the targeted visitors sign up to be notified. The cost: a domain plus a Carrd or Framer page. The interpretation: a 10 to 25 percent visitor-to-email conversion rate from qualified traffic indicates real interest. Below 5 percent indicates the positioning, audience, or problem isn't landing.
The smoke test ad. A small Google or Meta ad campaign ($50 to $200) pointing to the landing page. The signal: cost per email capture from cold traffic. The interpretation: under $3 per email indicates the product can plausibly be marketed profitably. Above $10 per email signals either weak positioning or a market that doesn't search for this problem in the way the ad assumes.
Customer interviews. Five to ten conversations with people in the target audience. Not a focus group, not a "would you use this" survey. Specific questions about what they currently do for this problem, what they've tried, what they spend, and what frustrates them. The signal: do they describe the problem in your words or do you have to translate? If they don't naturally describe the problem the product solves, the product is solving a problem they don't recognize.
Pre-orders or a paid waitlist. The strongest signal of all. A landing page that takes money (Stripe, $10 to $50 with a clear refund-on-fail promise) for early access to a product that doesn't exist yet. The signal: cash converts. People talk. People pay. Even five pre-orders is meaningful signal because no one pays for vapor unless they actually want it.
What "validated" actually means
Validation is a probability shift, not a guarantee. A well-validated idea has an 80 percent chance of being a viable product. A poorly validated idea has a 5 percent chance. The work is in moving from the 5 percent to the 80 percent range, not in achieving certainty.
Concrete validation thresholds for an indie product:
Landing page captures 50 to 200 emails from targeted traffic in two to four weeks. The "targeted" qualifier matters; 200 emails from a viral tweet is not the same signal as 200 emails from category-search traffic.
Smoke test produces a cost per email under $4 from a $100 to $200 ad spend. If the cost per email is much higher, the unit economics of acquisition will be hard later.
Customer interviews surface specific, repeatable pain points. Five out of ten interviewees describing the same problem in similar language is a strong signal. Five out of ten describing five different problems is a weak signal.
Pre-orders convert at 1 to 5 percent of landing page visitors. Even one pre-order from cold traffic is meaningful, and five to ten pre-orders is enough to justify building.
Hit at least three of these four thresholds before committing engineering time. Hit two of four and the idea is borderline; iterate on positioning before building. Hit one of four or none and the idea isn't ready.
The anti-patterns
Validation that doesn't validate. Five recurring failure modes.
"Would you use this" surveys. Affirmative responses are not validation. People say yes to be polite, and even a real "yes" doesn't predict behavior. The honest version of this question is "are you using something for this problem now, what is it, and what would make you switch."
Friend feedback. Friends are biased, polite, and unrepresentative. Friend validation has the highest false-positive rate of any method.
Vanity metrics on social media. A tweet about the idea getting 100 likes is not validation. Likes are free. The cost of attention vs. the cost of an email or a dollar is orders of magnitude different.
Domain-name vibes. Buying a domain and feeling excited about it. Not validation.
Talking to people who are already in your audience for unrelated reasons. Asking your Twitter followers (who follow you because you build stuff, not because they share the target audience of this specific product) skews heavily positive. The validation needs to come from the actual target audience, found independently.
The most common mistake
The biggest validation mistake isn't lazy validation. It's good validation followed by ignored results.
A founder runs a landing page test, captures 23 emails in three weeks against a 200-visitor goal, and proceeds to build the product anyway because "the positioning probably wasn't right." Sometimes that's accurate. Usually it isn't. The validation produced a signal and the signal was ignored.
A discipline that works: write down the threshold before running the test. "If this landing page captures fewer than 50 emails in three weeks, the idea is paused for a month." Pre-committing to thresholds reduces post-hoc rationalization. The threshold can still be revisited later, but the explicit "this didn't hit the bar" moment forces reconsideration rather than drift.
The decision after validation
Validation produces one of three outcomes.
Pass. Build the minimum viable version of the product. Use the email list captured during validation as the launch audience. Skip the "find first users" stage of post-launch because validation already produced them.
Fail. Either kill the idea or iterate on the positioning. Killing is the right call more often than founders accept. Iteration is appropriate only when the validation revealed a specific reason for the failure (the wrong audience, the wrong description, the wrong price), not a general "this didn't work."
Inconclusive. The signal isn't strong enough either way. The right move is usually to refine the validation (better targeted traffic, better interview pool, clearer landing page) rather than to declare ambiguous results as a pass.
What this means in practice
A reasonable validation sprint for an indie product takes three to four weeks and costs under $300 between the domain, the landing page tool, and a small ad spend. The output is a clear yes, no, or "iterate and re-run." That's three to four weeks of relatively unglamorous work that prevents months of building the wrong thing.
The makers who skip this stage do so because validation feels less productive than building. Building feels like progress. Validation feels like talking. The trap is that building the wrong thing is the most expensive form of unproductive activity available to an indie founder, and the only thing that prevents it is the kind of validation that produces uncomfortable signal early.
The makers who reliably ship products that people pay for don't have better ideas. They have a shorter feedback loop between idea and signal.
Plug Your Build is a permanent directory for indie makers across SaaS, newsletters, courses, Gumroad assets, Discord servers, and more. Standard listings start at $3.99/month and stay live indefinitely. Submit yours here.