Two directories, same premise, different age
twelve.tools and fazier.com both do the same thing: founders submit a link, the directory lists it, everyone hopes for a bump. We picked these two because they are structurally different enough to be a real comparison and old enough to have a track record worth measuring. fazier.com is the materially older, more established directory, stable from day one. twelve.tools took roughly a year to stop being sporadic. That gap matters for reading everything that follows: a directory that took a year to find its own footing should not be judged by whether its most recent members already have traction. They have not had the runway.
| twelve.tools | fazier.com | |
|---|---|---|
| First real Chrome traffic detected | Jan 2024 | Oct 2023 |
| Stable, continuous traffic since | Feb 2025 | Oct 2023 |
| Members audited | 517 | 567 |
Before trusting a single number below
Every one of the 1,084 members turned up independently in Common Crawl's own crawl of the web. 1,084 of 1,084, 100%. That is not automatic. Common Crawl does not index everything; crawl budget is earned by real link centrality. A dead domain, a squatted parking page, or a site that only 'exists' because a directory claims it does would very plausibly fail to show up here. None did. Every member checked in this audit is a genuine, externally-crawled site.
What you actually get when you're listed
We checked the real listing pages, not the homepage (which is mostly nav and social links), and isolated the actual visit-website button on each one. twelve.tools: 1,569 member links checked, 100% dofollow, zero exceptions. fazier.com: 1,546 member links checked, 89.7% dofollow, 10.3% nofollow. One fazier.com listing page even carried two different visit-website buttons with different rel attributes on the same page.
Open question: twelve.tools links its featured member on 93% of listing pages. fazier.com does it on only 33%. We do not yet know why. Could be real (many listings never added the link) or a naming-pattern miss in our extraction. Flagged, not resolved.
Where the members actually rank
Top 100,000 domains globally is Fortune-500-adjacent territory. Effectively no startup clears it, listed or not. We used thresholds an early-stage company could plausibly hit. About a quarter of each cohort cracks the top 5 million, which is genuinely meaningful authority for an early-stage company. About half of twelve.tools' members land in the top 10 million.
| Authority threshold | twelve.tools | fazier.com |
|---|---|---|
| Top 1,000,000 globally | 35 of 517 (6.8%) | 59 of 567 (10.4%) |
| Top 5,000,000 globally | 133 of 517 (25.8%) | 166 of 567 (29.3%) |
| Top 10,000,000 globally | 259 of 517 (50.2%) | 243 of 567 (42.9%) |
Do members actually go on to get traffic?
This is where we almost published the wrong story. First pass compared each year's traffic count against today's full member roster. That made early years look artificially weak: 7% in 2024, 28% in 2025, 55% in 2026, a tidy little growth curve. Except a member that joined in 2025 did not exist yet in 2024. It was not failing to get traffic. It was not in the race yet. We rebuilt the denominator around whether each member actually existed that year, checked directly against Common Crawl's page captures, not the traffic metric itself.
Correction, filed mid-audit: the apparent growth curve was a denominator error. Once a member exists, 63 to 83 percent clear Chrome's real-traffic floor almost immediately. The old numbers' apparent climb was just new members joining the roster, counted against a denominator that had not caught up yet.
| Year | twelve.tools | fazier.com |
|---|---|---|
| 2024 | 38 of 55 (69%) | 145 of 174 (83%) |
| 2025 | 146 of 192 (76%) | 231 of 319 (72%) |
| 2026 | 285 of 449 (63%) | 289 of 458 (63%) |
Denominator = members confirmed to exist that year via Common Crawl page captures, not today's full roster.
Do members link back, and why do some stop?
Checked directly in the link graph: does an edge exist from the member back to the directory. twelve.tools: 267 of 516 (52%) link back. fazier.com: 152 of 567 (27%). Real mutual engagement. We spot-checked one (saypi.ai): a genuine embedded image badge pulled from twelve.tools' own hosted assets. Not a stray mention. Current authority tier barely moves the needle on who keeps the badge. But when we split by whether a site already had traffic before it was listed, the gap opened wide.
The badge is training wheels. Great for getting your legs under you, and it comes off once you already know how to ride. Honest caveat: twelve.tools' 'already had traffic' bucket is only 8 members, thin. fazier.com's 46 is the number doing the real work here.
| Links back... | twelve.tools | fazier.com |
|---|---|---|
| ...already had traffic before listing | 1 of 8 (12.5%) | 3 of 46 (6.5%) |
| ...did not have traffic yet at listing | 267 of 509 (52.5%) | 149 of 518 (28.8%) |
The ~4x gap holds even restricted to members currently in the top 5M authority in both buckets, so it is not just a size effect.
What this does and doesn't prove
None of this is causal. These founders launched, marketed, and submitted themselves everywhere they could. The directory listing is one input among a dozen. We cannot say the listing caused the traffic. What it does kill is the flat 'directories do nothing' take. Most members that exist, 63% to 83% of them, go on to clear real, Google-measured Chrome traffic. That is the norm in both cohorts, not the exception. And the actual mechanics of what a directory gives you, link type, authority tier, whether members stick around and link back, are genuinely different between two directories that look identical on the landing page.
Method and data sources
Both datasets are public and free to query: Common Crawl's webgraph (who links to whom, what kind of link, whether a domain is real enough to earn crawl budget) and Google's Chrome UX Report via BigQuery (real Chrome traffic history, monthly, not a vanity snapshot). Member listing pages were enumerated via Common Crawl's CDX index. Member links were isolated from directory nav and social noise via slug-to-domain matching. A raw external-anchor count overstates member links by roughly 4x if you skip that step. Rel attributes were parsed from the actual WARC HTML. Member domains were joined against Common Crawl's PageRank panel and the real monthly CrUX tables, never the single-snapshot panel field, which undercounts badly. Extraction accuracy was checked against a genuinely random seeded sample: 30 member-matched listing pages per directory, re-fetched, page title checked against the extracted domain. 60 of 60 correct. The one confirmed slug-matching false positive across both cohorts was whop.com, excluded from all twelve.tools counts.