Audience research: where your people already are

Picking communities by follower count is how most launches fail. A large community that removes every product post is worth less than a small one where people ask for recommendations.

The useful question is not where the crowd is. It is where products like yours already get discussed, and whether someone in your position is tolerated there. GummySearch answered a close version of that question until it closed, and we mapped the alternatives its former users moved to.

Communities, sorted by what happens in them

• How many products the radar found being discussed there, which is the clearest sign of a community that talks about tools at all.
• How many distinct authors take part, so you can tell a busy room from a loud handful.
• Which category the community leans towards, which is how you find rooms adjacent to yours.
• Whether a founder can post there, read from what actually happens to product posts rather than from the sidebar rules.

That last point is the one people get wrong on their own. Written rules and enforced rules differ, and the gap is only visible in what survives.

When to post, not just where

For the communities in your plan, the hourly rhythm shows when their conversation actually happens, by hour of the week. It is computed from mention timestamps in a recent window rather than from general advice about Tuesday mornings.

The window is deliberately short. Older archived data has no reliable hour attached, and mixing it in would show a quiet stretch that is an artefact of the archive rather than of the community.

Learning from posts that worked

For a given product, playbooks show the posts its own makers wrote about it, with the original text. Not a summary, not an inferred best practice: what was actually published, and how it landed.

Reading three of those in your category teaches more about tone and framing than any generic guide on posting. The same material read against a named rival becomes competitor tracking rather than audience research, and the dashboard treats them as two views of one dataset.

Why community fit beats community size

A community of two million people that treats every product post as spam will remove yours within the hour, and nobody will have read it. A community of forty thousand where people ask for recommendations every week will read it and answer.

Size is easy to measure, which is why everyone uses it. Fit is harder, and it is the only one that changes the outcome. The counts here are built for fit: how many products get discussed, by how many different people, and what happens to product posts.

The same logic applies to adjacency. The room where your exact product is discussed may be small or hostile, while a neighbouring category tolerates makers and asks the same questions. Sorting communities by category rather than by keyword is what makes those rooms visible, and the same taxonomy drives the ranking of products a market recommends.

You can see that sorting at work without an account: the free subreddit finder ranks communities by product category and shows, for each one, how much of its product talk belongs to that category and whether links there usually arrive inside a post.