5 days after my "zero sales" post — traffic dropped as expected, but something interesting happened
Quick update on knallhart.dev since the last post got more engagement than I expected (thanks to everyone who commented, genuinely).
The numbers: traffic peaked around 33 visitors/day during the thread's active period, now down to about 10/day as the post ages — exactly the "single source of traffic" risk a few people warned me about. Still zero sales.
What changed based on feedback here:
- Added real screenshots of actual delivered emails to the landing
page (the "does the preview match the deliverable" gap a few
of you flagged) - Added a pull-quote from a real roast higher up, so the proof
isn't buried below the fold - Built a shareable result card — every roast now generates an
image meant for sharing, plus a pre-filled "share on X" link - Set up proper funnel tracking (scroll-to-form, checkbox,
checkout-click) instead of guessing where people drop off
The one genuinely new signal: two visits from Google search this week, the first organic search traffic since launch. Small, but it's the first channel that doesn't depend on me actively pushing something somewhere.
Reached out to two people from this thread directly for free test roasts in exchange for honest funnel feedback — no replies yet.
Not sure yet if this is a demand problem or a distribution problem. Leaning towards distribution, but trying not to just believe what's comfortable to believe.
Not rushing to explain away zero sales is genuinely hard discipline. The instinct after a post like that is to immediately start tweaking things to feel like you're doing something, but tweaking before understanding usually just adds noise to a signal that was already thin.
2 organic search visits is a small thing but it's the kind of small thing worth paying attention to, since it means something is starting to exist outside of attention you directly created. That's a different kind of signal than traffic you have to keep generating yourself.
Exactly the distinction I keep coming back to. Everything else on the list right now is something I have to actively push — a post, a comment, an ad spend. The two organic visits are the first sign that something exists independently of that effort, even if it's tiny. Trying to resist over-reading two visits into a trend, but it's the one number this week that didn't require me doing anything that day.
Respect the "trying not to believe what's comfortable" line — that instinct is the whole game.
Honest take: at ~10–33 visitors/day you can't actually answer demand vs distribution yet. ~40 anonymous visitors is statistical noise — zero sales from that tells you almost nothing, and polishing the funnel (the screenshots, pull-quote, and share card are all good moves) is optimizing a page that not enough of the right people have seen to judge.
What cuts through fastest isn't more traffic — it's 5–10 real conversations with people who fit your ICP. Watch them use it live, ask what they'd pay. Qualitative signal answers "is this a real problem" in a week; 40 anonymous visits never will.
And the free test-roasts probably got no replies because free + cold to a stranger is the weakest ask there is. If you go direct, make it warm and specific — "built this because [their exact problem], roasted your [thing], here's what it found" — a person, not a broadcast.
Distribution's likely the nearer bottleneck, but don't let zero sales from a tiny sample convince you the demand answer is in yet. Keep going.....
The "free + cold = weakest ask" line lands, and explains the silence better than I'd been telling myself. I asked permission first ("would you be open to...") instead of just doing it and showing up with the result — which is exactly the warm/specific version you're describing, just inverted into a weaker ask.
Noted on the sample size too. Easier to keep believing the funnel tweaks matter when there's no real data to contradict that yet — appreciate the push not to mistake polish for progress.
The traffic drop after a viral post hit is always such a gut punch, but that "something interesting" twist at the end is what I'm curious about — did it end up being a net positive somehow?
Honestly, the "interesting" part was just two organic search visits — first time anything's come in without me pushing it somewhere. Small enough that calling it a "twist" oversold it a bit in hindsight, but it was the one number that felt different from everything else that week.
The honesty here is refreshing — 'trying not to just believe what's comfortable to believe' is genuinely hard to practice. One way to start separating demand from distribution: look at where funnel drop-off happens before checkout. If people reach your CTA or pricing section and still bail, that points to pricing or trust gaps. If they bounce early before even scrolling to the form, it's more likely a messaging or demand issue. Your new funnel tracking should surface exactly this signal within a week or two. Also, two organic Google visits this early is actually a good sign — that traffic tends to be higher-intent than the community spike.
That's exactly the breakdown I've been trying to read, though the data's still thin. So far: people do scroll to the form (saw about 24% of visitors reach it on one recent day), but nobody's checked the consent box or clicked submit yet — which based on your framework points more toward pricing/trust than pure demand, assuming it holds with more volume. Too early to be confident, but it's the first time the funnel data has actually pointed somewhere specific instead of just "not enough sales yet." Good context on organic traffic being higher-intent too, hadn't thought about it that explicitly.
33 visitors a day dropping to 10 is too little traffic to separate a demand problem from a distribution one, and waiting for sales to answer it will take months. Read the funnel as rates instead: scroll-to-form, then form-to-checkout-click. If lots of people reach the form and nobody clicks checkout, that's an offer or pricing read you can get this week off 70 visitors instead of 700.
The shareable result card is the part I'd push hardest. A roast tool has something most products don't: the output itself can do the spreading, if it's sharp enough that the person roasted wants to post it. So I'd spend a cycle making the card funnier and meaner before chasing new channels, since that's a content-quality lever you fully control.
On the two Google visits, dig up the exact query in Search Console. If it's a "roast my X" type search, that's exactly the channel you said you want, the kind that compounds without you pushing it. A few pages built around the variations people search could turn those two visits into a real trickle.
The card point lands, that's actually a controllable lever rather than another "wait and see" channel. Going to spend real effort sharpening the wit on it before chasing anything new — funnier and meaner is a fair brief.
Good idea on Search Console, hadn't thought to check the actual query behind those two visits. Will dig that up. If it really is "roast my X" style searches, building a few pages around those variations sounds like the cheapest possible bet given it's the one channel that doesn't need me to push anything.
One thing I liked here is that you didn't rush to explain away the zero sales. It's easy to convince yourself it's "just distribution." Waiting until the funnel data tells a clearer story is a much stronger approach than optimizing based on assumptions.
Appreciate that. Tempting to rationalize it as "just distribution"
this early, but the data doesn't actually support either story yet,
so didn't feel right pretending otherwise.
That's the part I found interesting.
The question that stayed with me wasn't whether you'll eventually get sales.
It was how you'll decide what the first few sales actually prove when they do arrive.
Happy to explain what I mean if it's useful. What's the best email to reach you on?