Reading demand signal without a research budget
How do you validate a product idea for free? You don't need a research budget to find real demand signal. Most of it is already sitting in places a founder can look without paying for anything. The hard part was never access. It's reading what's there accurately, without your own hope or dread doing the reading for you.
Search: what people look for, and don't
What people actively search for is one of the most honest signals available, because it's unprompted: nobody's answering a survey question, they're just looking for something they already want. The absence of search volume around your specific framing is signal too, not nothing: it can mean the category is real but the language you're using for it isn't how buyers think about the problem yet.
Adjacency: what they already buy near your gap
Buyers rarely articulate a gap directly. They route around it: buying a workaround, a partial substitute, or two products stitched together to approximate what they actually want. Watching what gets bought adjacent to the gap you're filling is often a clearer demand signal than asking someone directly, because it's revealed by real spending instead of a hypothetical answer.
Complaints: the gap stated out loud
Reviews, forum threads, and support complaints about existing options are one of the few places buyers describe the gap in their own words, unprompted, usually while venting about something that failed them. That specificity is valuable: it tells you not just that demand exists, but exactly which part of the current options isn't working.
Competitor traction: demand someone else is already capturing
If a comparable product is getting real traction, that's demand evidence whether or not you compete with it directly. It confirms the category can convert, not just generate interest. The absence of any traction anywhere nearby is signal in the other direction, and worth taking seriously rather than explaining away.
Where doing this by hand hits its limit
All four sources are real and free. What they aren't, done by hand, is weighted or synthesized. A founder reading their own search results, adjacency, complaints, and competitor traction is also the person least able to grade that evidence objectively, for the same reason self-diagnosis is unreliable generally: hope and dread shape which signals get noticed and which get explained away. The sources aren't the hard part. Reading them without your own stake in the answer is.