Three weeks ago I posted here on Day 7: I am building Bersyn, an AI visibility tool for B2B SaaS, trying to get 20 paying customers in 40 days, sitting at 0. I said I would keep posting to stay honest. So here is the ugly middle.
Where I am: Day 34. Still 0 paying.
That is the headline and I am not going to dress it up. But the last three weeks taught me more than the first week did, and two strangers in the comments reframed the whole thing, so this one is worth writing down.
What I actually did
I went deep on one channel: scan-based cold email, but founder-direct. Instead of hitting hello@ and info@ and contact@ (which, as I learned on Day 7, is a black hole), I started finding founders' real personal emails from their own public GitHub commit history, and emailing them one at a time with the specific result from their own scan. "I asked ChatGPT, Claude, Gemini and Perplexity which tool in your category they would recommend. Here is who they named, and here is where you were invisible." About a dozen of those went out, every one to a verified founder, no shared inboxes.
The lessons that actually cost me something
1. The format was working against me. My first emails looked like marketing: a centered HTML column, a brand font, em dashes everywhere, and worst of all a "this message is from a mailing list" unsubscribe banner at the top. To a founder that screams automation. I tore it all down to plain text, left aligned, written the way I would actually type it to a peer. No unsubscribe header, because these are genuine one to one emails.
2. Honest beats clever. My first version pretended I had stumbled on their result while "poking at something." Sharp technical founders see through that instantly. The version that feels right just says it: "I am Gissur, I build Bersyn, I ran your product through it on purpose, here is what I found." Owning the vendor angle reads as more credible, not less.
3. I instrumented myself blind. This one stings. I was sending from one domain and the replies were landing in an inbox my own tooling could not read. For days I had no idea whether anyone had even responded. If you are automating outreach, make sure you can actually see the replies before you scale the sends.
The two comments that reframed everything
Someone named Harjot left a comment on the Day 7 post that I keep coming back to: zero paying with zero conversations is a distribution problem; zero paying with a lot of "interesting, but..." conversations is a positioning problem. They need completely different fixes, and conflating them wastes the clock.
That landed. I had been treating 0 customers as a conversion problem and grinding on email copy, when the truth is I just was not talking to enough people yet. I am in the distribution quadrant. The fix is not better copy. It is more of the right people seeing the thing.
And here is the part that made me feel a little stupid: my own data already told me what works. The single biggest traffic driver I have had is public scan teardowns, posts where I run the four AI models on a category and publish exactly who they recommend and who they ignore. One of those drove around 40% of a week's traffic. And what did I do? I quietly de-prioritized it and went all in on the slow, invisible channel.
So here is the change, starting today
Content teardowns become the primary channel, not a side quest. I am going to publish them regularly: pick a category, run all four models, show the receipts, name names. Cold email stays, but as a slow drip in the background, not the engine. And I am tracking the only leading number that matters right now, conversations per day. Today it is basically zero. That is the number I am trying to move, not the customer count.
The wall I still have not cracked
The honest open question is the same one from Day 7: people run the free scan, sit on the result, sometimes come back, and do not pay. Another commenter ran my problem through Gemini and the diagnosis was sharp enough that I am going to steal it. The free scan satisfies curiosity but does not create urgency. It shows the gap and the suggested fix at the same time, so people think "great, I will just do this myself." The proposed fixes: quantify the loss in actual money, gate the corrective content behind the paywall instead of giving it away, offer a cheap one time "fix your top 3 gaps" purchase instead of leading with a monthly subscription, and frame the whole thing as a leak in your funnel rather than a "visibility score." That is my next build, once there is enough traffic to test it on.
Day 34. 0 of 20. Still going, just pointed at a better channel now.
If you have gotten a free tool to convert to paid, what actually moved the needle: the result page copy, the follow up sequence, or changing the shape of the offer? Genuinely asking.
Top comments (0)