How to Avoid Building a Product Nobody Buys

There is a specific dread that arrives midway through building a product: the suspicion that you are polishing something nobody asked for. It usually shows up at night, it usually gets waved away with more building, and it is right more often than it is wrong.

The comforting version of this article would tell you the dread is impostor syndrome. This is not that article. Products that do not sell are mostly not bad products. They are products built inside one of a handful of traps, and every one of those traps was visible before the build started. Here are the five that do most of the damage, each paired with the cheap check that catches it while the fix still costs nothing.

Products fail before they are built, not after

When a product does not sell, the launch gets the blame: wrong day, wrong channel, weak announcement post. Rerun the launch, the thinking goes, and the result will change. It almost never does, because the result was decided earlier. A problem nobody pays to solve was chosen, or nobody was told during the months of building, or the product was shaped for a buyer who exists only in the founder's head. The launch did not fail. The launch reported the failure.

The useful consequence: decisions made before the build can be checked before the build, and the checks are cheap. There is a complete validation system for doing this thoroughly, ending in a written go or no-go decision, in how to validate a business idea before building it. What follows is the minimum version for people who will not run the full system: five traps, five checks, one page.

Trap one: building in private (the check: ten conversations)

The most common trap feels like virtue. Heads down, no talking, ship the thing, reveal it when it is worthy. Meanwhile every judgment about what the buyer needs, which sections matter, what the product should even be, is a guess, and the guesses compound weekly with nobody to catch them.

The check: message ten people who plausibly have the problem and ask a question, not a pitch. "What is the most annoying part of this for you?" works because it invites detail and asks for nothing. The replies steer the build. Silence from all ten is also an answer, about whether you can actually reach the person you imagine buying.

The check works mid-build too. If you are weeks in and have told nobody, the ten conversations are this week's work, and whatever they change about the product is cheaper to change now than after launch.

Trap two: trusting compliments (the check: count actions, not opinions)

"Great idea, I would totally buy that" is the most dangerous sentence in early-stage building. It costs the speaker nothing, it is almost always kindness rather than commitment, and it accumulates into a ledger of encouragement that feels like demand. Then the product ships and none of the compliments convert, because none of them were ever purchase decisions.

The check: keep score of actions only. A pre-order. An email address handed over for a waitlist. A long reply describing someone's actual situation. An unprompted "tell me when it is ready." Each of those cost the giver something, money, contact details, effort, and that cost is what makes them evidence. Opinions are excluded from the count on principle, however warm.

The honest test is one question: what has anyone done, that cost them something, because this product might exist? If the answer is nothing, you do not have demand yet. You have politeness.

Trap three: the empty-niche fallacy (the check: look for busy shelves)

"There is nothing like this out there" feels like a discovery. It is usually a verdict. Markets are crowded wherever money is, so a shelf with no products on it most often means the people with this problem do not pay to solve it. The absence of competitors is the market pricing the idea at zero.

The check inverts the instinct: go looking for competitors, and be relieved when you find them. Search the marketplaces where you would sell. A busy shelf, products with visible sales and reviews, where every product misses something, is the sign you want: proven spending plus room to be better. The research method for reading a shelf properly is in how to find a profitable digital product niche.

If a genuinely thorough search finds nothing anywhere, no products, no paid tools, no services, treat that as a red flag demanding extraordinary evidence, not as an opportunity nobody else was smart enough to see.

Trap four: scope creep as procrastination (the check: one document, one outcome)

Somewhere mid-build, the product starts growing. A bonus module. A video version. Three extra chapters for edge cases. It feels like adding value, and sometimes it is. More often it is fear wearing a productivity costume: every week spent expanding scope is a week nobody can tell you no, and the reveal you are dreading gets postponed with a clean conscience.

The check: write one sentence naming one buyer and one outcome, then ask whether one core document could deliver that outcome. If the honest answer is yes and the current scope is triple that, the extra scope is not serving the buyer. Ask what it is protecting you from.

A version one that ships this month and hears a no is worth more than a deluxe version that hears the same no in six months. The no is the valuable part. It arrives with reasons attached, and reasons are what version two is made of.

Trap five: launching to nobody (the check: a launch list before a launch)

The fifth trap detonates on launch day but is armed much earlier. The product is done, the page is live, and there is no one to tell. So the founder posts twice into the void, watches nothing happen, and concludes the product failed. What actually failed is distribution that was never built. The product never got tested at all.

The check: before the build finishes, write an actual list of names. People you talked to while validating. People who reacted when you mentioned the problem. Communities where a genuinely useful post would be welcome. Names with context, not follower counts. The full operating procedure, who goes on the list, what the message says, what to do with silence, is in the first 25 people to tell about your product.

If you are mid-build and the list would be empty, notice that this is trap one wearing different clothes, and the fix is the same: start the conversations now, while there is still time for the people in them to end up on the list.

The one-page pre-build audit

Five traps, five checks, one page. Run it before the build starts, or today if the build is already running.

• Ten people with the problem have been asked about it, and some replied with detail
• At least one person took an action that cost them something: a pre-order, an email address, a detailed reply, an unprompted "tell me when"
• Competitors exist and visibly sell, and you can say what each one misses
• The product is one buyer, one outcome, one core document
• A written launch list of real names exists, each with context for why they would care

An unchecked box is not a verdict on the idea. It is this week's work, named precisely. Every one of these is cheap to fix before launch and expensive to discover after, which is the entire argument for checking now.