August 24, 2026 · Clemens

What we learned launching on DevHunt: 3,351 impressions, 22 visitors, zero signups

The DevHunt front page tells you Daytona got 200k+ impressions. DirectoryKit and Pulsetic both over 125k. More than 100,000 developers have found tools there. We read all that, saw the listing cost $49, and decided it was cheap enough that we didn't need to think about it very hard.

Scroll up from those numbers instead of down, to the week that's actually running, and the picture changes. The week we launched Context Goblin into was led by a tool with 3,345 impressions and seventeen votes. Seventeen. We finished at the top of that week in second place, on fourteen votes and 3,351 impressions.

Placing at the top of the week sent us 22 visitors and zero signups. This post is about the distance between those two sets of numbers, and about the minute of checking that would have told us to expect it.

Nobody lied to us. The past winners are clearly labeled as past winners. But you form your sense of how busy a place is from the page you land on, and that page is mostly history.

What the current week looked like

Open the board, ignore the leaderboard of all time, and read the vote counts at the top of the week that's live. That's the whole check.

Doing it would have shown us a week where seventeen votes wins, where the tools below the leader run from roughly 800 to 2,300 impressions, and where our listing was going to land somewhere in that band no matter how good the tagline was. The question then becomes an easy one. Is a channel that tops out around three thousand impressions worth a launch slot and the attention that goes with it?

Maybe still yes at $49. But you'd be answering the right question, and we weren't. We answered "is this site legitimate and does it reach developers", which it is and it does, then let the six-figure numbers on the page fill in the part about volume.

What finishing second bought

Six days in, our listing shows 14 upvotes, 3,351 impressions, and DevHunt's own #2 Week rank label. Our analytics recorded 22 sessions carrying a DevHunt referrer.

Those two figures count different events, which is worth saying plainly. An impression is our card being rendered on a page. It counts whether the person read the tagline, scrolled past on the way to the tool below, or left the tab open while they made coffee. A session is someone who clicked, waited for our page to load, and looked at what we do. Only one of them can turn into a customer, and every directory has a reason to lead with the other, because it's the number that grows with their traffic rather than with your product.

Nothing went wrong mechanically. The listing was live, the link worked, and people clicked it at about the rate you'd expect. 22 out of 3,351 is 0.66%, unremarkable in both directions.

None of the 22 signed up.

What 22 visitors is worth

A landing page converting cold traffic at 2% to 3% is doing well. Ours ends by asking for a GitHub App install on a repository you care about, so realistically it's lower. Run 22 visitors through even a generous 3% and the expected result is 0.66 signups.

That's the whole explanation. 22 visitors produces zero signups most of the time, whatever the page says and whoever wrote it. The sample was too small to tell us anything about the product, which is more irritating than an outright failure, because a failure at least teaches you something.

And we placed second. Second is close to the ceiling of this channel in this week. If finishing at the top of the week produces 22 clicks, the whole range of outcomes available to us was small.

What we learned

Read the current week, not the hall of fame. The front page shows what a directory has ever done. You're buying what it does next week. Those differed by a factor of fifty here, and the difference was visible in under a minute once we went looking.

Decide the number before you pay. We never wrote down what success meant, so the leaderboard position became the definition of success by default and stayed that way until someone opened the analytics. "This works if it produces five signups", written on the way in, would have told us where we stood on day two instead of day six.

Directories are not distribution. The channels that have produced actual conversations for us are the slow ones: talking to teams who have the specific cross-repo problem we built for, and writing things worth reading. Neither has a leaderboard.

If you're about to launch on one

Overthink it. At $49 it's reasonable, but for us it didn't buy more than a minimal SEO boost for our landing page. If you do, size the expectation to the week you're launching into rather than to the numbers on the front page. Write down the number that makes it worth repeating, and check your own analytics rather than theirs.