I run a launch directory. Every Sunday for the last 14 weeks I have published a recap with the week's raw numbers in it: how many products were submitted, how many went live, how many upvotes were cast, how many builders joined. I did that for transparency, not for analysis. Then I lined all 14 weeks up next to each other and found that most of what I believed about launch day was wrong. Here is the short version. Across 14 weeks, 1,032 products went live and collected 7,636 upvotes between them. The best predictor of how a launch performs is not what the maker paid, not the category, and not the day of the week. It is whether the maker brought an audience with them. Paid placement buys position on the page. It does not buy votes. This piece is the data behind that claim, including the parts that make my own pricing page look bad. The dataset Nick Launches went live in May 2026. Every week since, a recap post has published the same five counters. Nothing here is modelled or sampled. It is a census of what happened. Week ending Went live Upvotes Live products DR --------------- May 10 70 280 76 1 May 17 85 463 158 May 24 62 493 260 May 31 54 268 310 Jun 7 55 287 364 54 Jun 14 56 247 462 54 Jun 21 79 378 601 57 Jun 28 75 668 667 60 Jul 5 77 365 744 61 Jul 12 79 568 819 63 Jul 19 85 919 908 65 Jul 26 82 881 1,063 68 Aug 2 84 855 1,144 69 Aug 9 89 964 1,202 71 Two caveats before the findings, because they limit what you should take from this. Submission counts were only added to the recap in week 11, so the first ten weeks have no denominator. And "live products" counts everything in the directory, which includes listings that entered outside a weekly launch window, so it does not reconcile line by line with the launches column. Where a claim below depends on a number I do not have, I say so instead of estimating it. Finding 1: the free launches keep winning Week 32 closed with 89 launches on the board. The winner was Lutains, a free launch, on 100 upvotes. Second place had 25. That is a four to one gap between first and second, and the product on top paid nothing. Twelve of the thirteen paid placements on that week's featured board finished below it. This was not a fluke week. In week 31 the top free launch finished on 24 upvotes, ahead of eleven of the thirteen paid placements. The pattern has held every week I have checked: a paid slot reliably puts a product in front of more people, and reliably fails to convert that exposure into votes on its own. The mechanism is not mysterious once you look at who wins. Lutains is a service where two people read your landing page by hand and send back an annotated file. The maker had a following that cared. The votes came from outside the directory, not from inside it. So the honest version of what a paid placement does: it holds your listing in a visible position for a fixed window. It is a distribution surface. It is not demand. If you have nowhere to announce your launch, buying position will not manufacture the announcement. Finding 2: the audience grew, the supply did not Launches per week have barely moved. The range across 14 weeks is 54 to 89, with no trend. Upvotes per week went from 280 to 964. Split it in half and the effect is clean. In weeks 1 through 10, launches averaged 5.8 upvotes each. In weeks 11 through 14, they averaged 10.6. The same volume of products, roughly twice the response. This matters for timing, and it points the opposite way to the usual advice. Builders often try to launch into a quiet week to avoid competition. On the evidence, competition for attention between launches is not the binding constraint. The size of the audience showing up is. A crowded week on a directory that 2,980 builders have joined is worth more than an empty week was in May. If you are sitting on a finished product waiting for a good moment, the good moment is the one where the most people are around. Finding 3: launch day is not one day The recaps track products across weeks, and the same names keep reappearing with higher numbers. One product went from 44 upvotes in its first week to 126 in its second, and held the top of the board both weeks. Another has been on the board three weeks running. Several carry a note in the recap saying they are still climbing from the previous week. Launch day framing is a habit borrowed from platforms where the page resets at midnight and yesterday's launches disappear. On a directory where the listing is permanent, votes keep arriving. The products that compound are the ones whose makers keep pointing at the listing after the launch window closes, in a changelog, a newsletter, a reply to someone asking for exactly that tool. The corollary is that a quiet launch day is recoverable. It is not a verdict. Finding 4: the queue is the real constraint Here is the number that surprised me most. In week 14, 238 products were submitted. 89 went live. The week 14 recap on Nick Launches, showing 238 products submitted and 89 live That gap is not a rejection rate, and I want to be precise about it because it would be easy to misread. Free launches cannot claim the current week, and there is a cap of 70 free slots per week. Paid placements are uncapped and publish on approval. So a submission arriving on Tuesday is usually scheduling itself into a week that has not happened yet. The backlog is a queue, not a reject pile. The practical consequence for you is a calendar one. If your launch needs to land near a funding announcement, a conference, or a Product Hunt run, submitting the week before is too late on most directories with a weekly cadence, mine included. Two to three weeks of lead time costs you nothing and buys you the week you actually wanted. Of the 89 that went live in week 14, 11 were premium placements. The other 78 were free. The free tier is not a token gesture on a board like this, it is the board. Finding 5: the listing appreciates whether or not the launch does The domain rating column is the one I did not expect to be interesting. It went from 1 in May to 71 in August, gaining nine points in the final month alone. A permanent listing with a do-follow link is worth more today than the identical listing was worth in June, and the maker did nothing to earn that. It came from the directory growing. This reframes what a launch on any young directory is. The upvote count is the visible part and it decays in attention terms within a week. The listing is the part that keeps working, and on a domain that is still climbing, it quietly gets better. When people ask me whether it is worth listing on a small directory, this is the actual argument, and it has nothing to do with launch day. What I would tell you to do with this Bring your own first twenty votes. Every winning launch in this dataset had an audience before it had a listing. A directory is a place to point people, not a substitute for having people to point. Submit two to three weeks early. The queue is the constraint, not the review. Treat the listing as the asset and the launch as the announcement. Attention fades within the week. The listing does not. Do not buy placement to fix a distribution problem. Buy it when you already have an audience and want the listing held in view while you point them at it. That is the case where it works, and it is the only case I can defend from this data. Where the numbers came from Every figure in this piece is published. The weekly recaps on Nick Launches carry the counters as they were recorded that week, including the week 14 recap with the full breakdown, and the ranked board for any week is public at the week URL, so the leaderboard screenshot above can be checked against the live page. I will keep publishing them weekly. If a finding here stops holding, that will be visible in the same place, which is the point of putting the raw counters in public in the first place.
