Every "launch your side project" checklist has the same list on it. Product Hunt. AlternativeTo. BetaList. There's An AI For That. Toolify. Futurepedia. G2 if you're feeling corporate.
I run a small free photo-filter tool. Before spending a weekend submitting to all of them, I did something I should have done first: I pulled the backlink profiles of the competitors who are beating me, and looked at where their links actually come from.
| site | total backlinks | what I sampled | | --- | --- | --- | | competitor A (direct rival) | 417 | 333 (80%) | | competitor B (much bigger, older) | ~13,000 | top 356 follow links by page authority | | competitor C (new, same idea) | 57 | 46 | | my own site | 38 | 36 |
I pulled these from Semrush's backlink report, paginating the table and scraping each row into a flat list: source URL, page authority, follow/nofollow, first seen. Then I grouped by domain and sorted.
Sampling caveat up front, because it matters for how much you should trust this: for competitor B I only looked at 356 of its ~7,100 follow links, about 5%. That's not a random 5%. It's the highest-authority 5%. Directory listings tend to be high-authority pages, so if they existed I'd expect them near the top. Their absence there is meaningful, but it's an inference, not a census. For competitor A I have 80% of the profile, which is solid.
I grepped both profiles for every directory name from the standard checklist: altern, producthunt, saashub, toolify, futurepedia, theresanaiforthat, betalist, slant, capterra, g2., stackshare.
204 links from 185 different personal websites, nearly all dofollow. Mostly Neocities and Nekoweb sites: the free-static-hosting corner of the web where people still hand-write HTML and keep a page called /links or /resources or /cool-sites.
One single Neocities site linked competitor B 168 times (it's a big personal directory, one link per tool it catalogues).
The striking part: those 185 sites barely overlap. Exactly one of them turned up in more than one competitor's profile. These aren't a network. They're 185 individual people who each found a tool and added it to their own list.
There's a second tier that's smaller but real: Substack newsletters doing "18 free tools for X" roundups, niche wikis, a couple of .edu pages where a lecturer keeps a digital-art tool list, and, in competitor B's case, a cluster of Chinese design-navigation sites that accept submissions.
My assumption was that "free web tool" is a category that lives in tool directories. For SaaS, maybe. For a thing you use once in a browser and never sign up for, the people who link to you are the people who collect things like that. And the people who collect things like that, in this niche, keep personal homepages.
That's not a growth hack. It's just: your links come from wherever your actual audience hangs out, and that's a different place for every niche. The generic checklist encodes someone else's niche.
Something else fell out of the data that's worth flagging if you ever run this check.
Competitor A's profile contains 122 links from a network of near-identical junk domains: beer-brand names, .xyz TLDs, all serving the same template. That's a third of the 333 links I pulled. My own tiny profile has the same shape: of 38 backlinks, 36 are auto-generated "Boost your Google rankings!" scraper pages that also point at everyone else in the niche.
I want to be careful here, because it's easy to read that as an accusation. You cannot tell from a backlink report who put a link there. Nobody asks permission to link to you. Scraper networks and spam farms hoover up domains and link at them automatically; showing up in one says approximately nothing about the site being linked. I'd assume mine arrived exactly the way theirs did. By robot.
The practical takeaway is just: when you read a profile, filter the junk before you count anything. Competitor A looks like a 417-link site. Strip the template spam out of what I sampled and roughly 210 links are left. And don't panic-disavow. A handful of obvious template spam is noise every site accumulates.
First, I stopped planning the directory sprint. Not because directories are bad, but because in this niche the evidence says they pay nothing.
Second, I made an actual site in that community: a small hand-written page about why old digital camera photos look the way they do, with the slider values I use, hosted on Nekoweb like everyone else there. Not a landing page. That crowd can smell a landing page from orbit.