Social listening for product validation

The cheapest validation is finding out whether strangers already describe your idea as a problem they have. If people ask for it in public, you have a market and a list of first conversations, and if two weeks of sweeps come back empty, you learned that for a few credits. Oetzi compiles the idea into searches on seven platforms and scores what comes back.

What does demand for an unbuilt product look like?

It looks like a workaround. On Reddit: "I built a spreadsheet for this because nothing does it properly". On Hacker News: "we wrote a script for this internally, why is there no product". On X: "why is there still no tool that does X, I would pay tomorrow".

The workaround is the proof. Someone spent their own time on the problem, which beats agreeing it is annoying. Count workarounds, not sympathy.

How do I describe an unbuilt idea in one sentence?

Describe the pain in the words people use, not the product you plan to build. Something like "people describing a manual process for keeping two systems in sync and asking whether a tool exists". That becomes native searches per platform plus a rubric, and seven days backfill immediately.

Write two or three versions using different vocabulary. Your category name will not appear, because the category does not exist yet. Plain english audience covers writing in the market's language.

How do intent scores read when you are validating?

Each match is scored 0 to 100 with the reason written out and tagged buyer, current customer, feature request, or competitor user. Read the distribution, not any single match. High scores with buyer tags mean people are looking and nothing serves them. Many low scores mean the pain is real but nobody is shopping, so it is not expensive enough yet. Competitor user tags mean something already exists, which changes the pitch from invention to replacement. An empty inbox is the clearest result of the three.

What do I do with the matches?

Talk to them. Reply in the thread, ask what they tried, and say you are looking into building something. Two reply variants are drafted in your voice and nothing posts without approval. The goal is conversations, not signups.

Save the phrasings. The words people use here become your landing page copy and the buying questions you track later. When you ship, finding beta users runs off the same signal.

What does validating an idea cost in credits?

Less than anything else you will try. Reading is free, and the free tier gives 10 credits with no card that never expire. A sweep of one signal across every platform is 2 credits and refunds when nothing qualifies, so 10 credits is five sweeps. On the $49 plan with 100 credits included, three idea signals swept twice a week is about 48 credits a month. Explore, the inbox, and translation are free, so a market arguing in three languages costs the same as one that does not.

Questions

How do I validate a product idea with social listening?

Write the problem as a sentence in the market's own vocabulary, sweep it across X, Reddit, LinkedIn, Threads, Bluesky, Hacker News, and Substack Notes, and read what comes back. Seven days backfill immediately. Count posts where someone built a workaround, because paying with their own time is the best evidence they would pay money.

What if the sweeps come back empty?

Change the sentence first, twice, using different vocabulary, because an empty inbox often means you described the idea in your words rather than theirs. If rewrites still return nothing, treat that as a result. Sweeps refund when nothing qualifies, so the test cost very little, and knowing early is the point.

Does an empty result mean the idea is bad?

Not necessarily. It means nobody is asking in public on these seven platforms, which is a real constraint if your plan was to find first users there. Some markets do their asking in private communities, at events, or inside vendor relationships. If that is yours, you need a different first channel, and it is better to know before you build.