Last week I wrote about sending my startup's launch press release to seven free sites.
https://hisashi.space/blog/free-press-release-sites/
That post was about getting it out. This one is about the five days after. I launched Pinateca, a project management tool, on September 24, and the English release went out through EIN Presswire that evening.
Reprints on about 800 sites, 3 visitors from them
EIN's distribution report looked great. AP News, around 250 USA TODAY Network local papers, about 19 TV station sites, a few hundred other local and lifestyle sites, 145 journalist contacts in Japan and 211 in the US. It was in Google News within an hour.
Then I opened GA4 for September 24 to 28 and filtered sessions by source.
・einpresswire.com: 2
・einnews.com: 1
・AP News, the local papers, the TV sites: 0
For comparison, my Japanese release on a small domestic site brought 10 sessions in the same window. The reprints are pages that partner sites generate automatically. Nobody wrote them and, as far as I can tell, nobody reads them.
What Ahrefs showed
I checked pinateca.com with Ahrefs' free authority checker four days after the release.
・605 backlinks
・554 linking sites
・4% dofollow
Almost every reprint marks its link as nofollow, which means search engines are told not to count it. Domain Rating stayed at 4.3 out of 100. So you do get a lot of pages with your name on them, you just should not expect an SEO bump from the links.
The first pitch arrived 44 minutes later
The release went live at 13:10 UTC. At 13:54 I got an email with the subject "An Interview request". The last line said there's a nominal fee for promoting the story.
Over the next five days:
・a UK wire offering to put my release on 65 regional newspaper sites (price after they see it)
・a magazine publisher who'd already written a draft feature about me, discounted if I paid within 48 hours
・another wire promising Yahoo Finance and Business Insider
・an agency asking whether ChatGPT recommends my product (it does, more on that below)
・a shared spreadsheet titled "$50 Each post sites", sent to me and 14 other companies at once
・EIN itself, twice, asking me to buy the next release
None of these is a scam exactly. But they're all written like editorial outreach ("interview", "feature", "we'd love to cover you"), and the price is always in the last paragraph or the second email. If a stranger found you through a wire service, scroll to the bottom before you get excited.
For what it's worth, this doesn't happen to everything I ship. Earlier in September I sent releases for two Mac apps, SpaceDeck and Koffret, through the same kind of free Japanese services, and got nothing back but the listing confirmations. Pinateca's Japanese release, on the day it went out, brought the EIN offer, an affiliate network listed on the Tokyo Stock Exchange, and an influencer agency. My read is that the people who sell to companies watch these feeds for B2B tools. A Mac app someone buys for themselves doesn't look like a customer with a budget. A SaaS billed per seat does. (I also used one extra Japanese site for Pinateca, so it's not a perfectly clean comparison.)
The one free offer
One pitch didn't ask for money. A site called CB Herald wanted to republish the release for free. I said yes, sent the Word file they asked for, and it went up four days later with a normal-looking article page and a dofollow link.
I also looked at the site itself. It has 7,039 posts, and 6,957 of them are under the same admin account. It is not in the Tranco top million. It is a reprint site. I am glad of the link, and I do not expect a single visitor from it.
Where the signups actually came from
Six people signed up on their own in those five days. I traced each one through the server logs.
・ChatGPT: 3 (Japan, Israel, Poland)
・Indie Hackers: 1
・my Facebook post: 1
・my newsletter: 1
・the ~800 reprint sites: 0
The ChatGPT ones are easy to spot because ChatGPT adds utm_source=chatgpt.com to the links in its answers. For the Polish signup there's an extra detail: 16 minutes before they landed, my logs show a request with the ChatGPT-User user agent fetching my home page. That is ChatGPT reading the page live to answer someone's question. A commenter on Indie Hackers pointed me at that trick, and it is the closest I have come to seeing the moment an AI recommends you.
Google's AI answers were citing me too, and not because of the release. Search Console shows my pages appearing in AI features from September 13, eleven days before launch day. The pages being cited were blog posts comparing competitors' free plans. One of them ranks 2nd to 3rd for "clickup free plan limitations" in regular search as well.
What I'd do after launch next time
In the order I'd do it:
1. Send one free wire release, and expect pages, not people. If you get a free offer from EIN, take it. It costs an afternoon and gets your name onto hundreds of pages, which can only help an AI that is trying to work out whether you exist. Read the terms first, though. openPR's terms say a free release can't also go out through another service, so I didn't send it there.
2. Write the posts people are already searching for. Not "we launched" posts. Posts that answer questions your buyers type before they've heard of you. For me that was "what's the catch with ClickUp's free plan" and "how many people does Asana's free plan allow". Those got cited by Google's AI answers before the launch did anything.
3. Get listed where your users already are. I cannot prove this one yet. Pinateca went live in the Zapier app directory on September 18 and in Pabbly Connect on September 28, and no signup traces back to either so far. I'd still do it first thing, since people browse a directory when they're actually looking for a tool, which is more than anyone does on a reprint page.
4. Watch your logs for AI traffic. A grep for utm_source=chatgpt.com and ChatGPT-User is enough to start. If your marketing pages are cached at a CDN, those fetches can stop at the edge and never show up in your server logs. My home page is not cached, which is why I caught that one fetch, and it's also why I know I'm missing the others.
5. Set a rule for the inbox before it fills up. Mine is simple. Anything that found me through a wire service and calls itself an interview or a feature gets read to the last paragraph before I reply.
Next time
I'd send it again as long as it's free. I just won't read the distribution report as a result anymore.
If a paying customer ever came to you from a wire release, which site was it on?
Top comments (14)
Thanks for publishing the real numbers, ~800 reprints for 3 visitors will save people an afternoon. Since Pinateca is a Mac app for individuals, who are you picturing as the first 10 users: freelancers, students, or people leaving a bigger tool?
Glad the numbers help! Small correction though: Pinateca is actually a web app for teams, billed per seat. The Mac apps in the post are SpaceDeck and Koffret. So of your three, it's the last one: people leaving a bigger tool. I built it because I was on Trello myself and didn't like how the bill jumps once a team grows past a certain size. My consulting clients' teams already run their projects on it every day, so they were the first users.
Thanks for the correction, I had Pinateca mixed up with the Mac apps. Consulting clients already running their projects on it is a strong start.
One thing that might help find the next teams: Trello bills a guest who's on two or more boards as a full paid seat (support.atlassian.com/trello/docs/...). Teams that work with outside people hit that fast, and their admins ask about it in the Trello questions forum on the Atlassian Community.
For your clients, was it that kind of jump (guests counted as seats), or just the team getting bigger?
It was the team getting bigger. My clients were also paying for a bunch of other subscriptions like email services and AI tools, so they didn't have room to add more cost for task management. I liked using Trello myself too, but I figured building my own would help me as well, so I did. I still use it every day and it's really nice to use.
That makes sense. Clients who already pay for email and AI tools don't want one more bill that grows every time they add someone, and you saw that from the inside on Trello.
That's probably your strongest line too. If it's not already on your pricing page, show what a team of 10 or 20 pays per month on Pinateca, next to what the same team would pay on Trello. An admin who's about to add five people is doing that math anyway.
I'm building FounderPulse, which looks for people who just posted about the problem you solve (here, team admins saying their Trello or Asana bill keeps growing) and drafts a first reply. There's a free test on founder-pulse.com, no signup. If you try it with pinateca.com, I'd like to know if the threads it finds are ones you'd actually answer.
Thanks for the tip! I actually have that on the compare pages already. For example, a team of 10 comes to $50 a month on Pinateca, versus about $110 on Asana Starter. Good luck with FounderPulse ☺️
Nice, $50 vs about $110 is a strong line. Here's where it answers someone directly, all on the Asana forum:
A plain answer to their question first, then the math if they ask. Thanks, and good luck with Pinateca too.
Thanks for digging these up ☺️
Looking at Asana's forum rules though, it seems they say you shouldn't promote your own product. Do you think it'd be okay?
Good catch. If their rules say no self-promotion, I wouldn't post Pinateca there: it gets removed and can cost you the account.
What still works:
Thanks, that makes sense. Answering their question first and keeping Pinateca in my profile sounds like the right way to do it on that forum. And the "free plan is now 2 seats" page idea is a good one ☺️
Glad it helps. If you write that page, send it over, happy to give it a quick read.
Thanks, that's really kind of you ☺️
The useful part of the five-day trace is that you stopped treating the wire's distribution report as a result. Since you've already decided to repeat a free release, the next distinction seems to be cost versus evidence: which one of the sources you traced deserves a focused second test, and what would that test need to show?
Good question. The wire didn't bring people directly, five days made that clear. But I wouldn't call it useless. ChatGPT brought 3 of the 6 sign-ups that week, and for one of them my logs show ChatGPT fetching my home page 16 minutes before they landed. If those hundreds of reprint pages are part of how ChatGPT got to know us, the release did its job, just not where the report says. So the second test I'd run is about timing: do ChatGPT sign-ups pick up in the weeks after a release, compared to weeks without one?