What's your actual cost per customer when you count your own time?

Yeah that matches what Im seeing. The audience temperature thing is real. When I post in places where people are already aware of the problem, a single reply can do more than an hour of cold posting somewhere broader.
But it also kind of proves the point about counting time. Finding those warm pockets is itself work. Lurking, reading threads, figuring out which subs or spaces actually have the right people. That research time never shows up anywhere but it directly determines whether your next hour of marketing is cheap or expensive.
So I guess my honest answer to my own question is that time CAC isnt really one number. Its two. The hours you spend talking to the wrong people, and the hours you spend finding the right ones. The first kind is what makes founders quit. The second kind is what actually compounds.

The $0.02 API cost versus the real cost when you count your time is exactly the gap most founders never quantify. And you are right that at pre-revenue the number is technically infinite, which is why most people avoid the math entirely.
To your question about when it becomes sustainable, I think the answer is when the ratio between new work and compounding work flips. Right now you are probably spending 90% of your time on things that produce a one-time result and 10% on things that keep working after you stop. SEO content is in that second category. A landing page iteration is in the first category unless you are testing something that permanently improves conversion.
For my project the shift started becoming visible around week 4-5. The content I wrote in week one was still generating visits without any additional time from me. Small numbers, but the curve was there. The cost per customer does not become sustainable at a specific revenue number. It becomes sustainable when the denominator starts growing faster than the numerator.
The SEO content you are building for your cover letter tool is probably the highest leverage work you are doing right now even if it feels slow. Every article that ranks is a customer acquisition channel that costs zero to maintain. How are you choosing which keywords to target?

Honest answer: I am still in the early stage. About 5 weeks in. But the traction signals are starting to show.
First two weeks were basically zero. Writing content, setting up platforms, publishing into the void. No organic traffic, no referrals, nothing.
Week three, Google started indexing pages. A few organic visitors trickled in. Quora answers started getting upvotes and clicks.
By week five, organic search traffic nearly doubled from the previous week. Still small numbers, but the direction is clear. Referral traffic from Quora and Medium has a 72% engagement rate, which means the people who do arrive are actually reading and using the tool.
The thing nobody tells you is that traction does not arrive as a moment. It arrives as a slope. You do not wake up one day with traffic. You notice that this week was slightly better than last week, and last week was slightly better than the one before. The compounding is real but it is invisible in the first month.
If you are building something, the main thing I would say is do not evaluate results before week 8-12. Everything before that is planting, not harvesting.

The ownership point is valid. Building on rented platforms is a real risk and I have seen enough Medium and Twitter horror stories to take it seriously.
That said, I think the "stop measuring time" part is dangerous advice for early stage founders. You should absolutely measure your time, not to calculate an hourly rate, but to understand which hours are building owned assets and which are just keeping you busy. Without that data you cannot know whether your system is actually compounding or just feeling productive.
I use Medium as a feeder channel, not a foundation. Every article there links back to my own site where the diagnostic tool lives. If Medium disappeared tomorrow, I would lose a referral channel but not the business. The site, the content, the SEO authority, all of that is owned.
The real question is not time versus ownership. It is whether each hour you invest increases the value of something you own. If yes, the math works. If no, it does not matter how sovereign your infrastructure is.

You are right, $30/hour is generous to myself. If I priced the same work at freelancer rates it would be $75-$100/hour and the CAC would be genuinely painful to look at.
But that is kind of the point of the question. Most founders use a comfortable number because the honest number makes the whole project feel irrational in month one.
To your second point, that is exactly the bet. Right now every hour is manual. But the content I wrote three weeks ago is still generating visits today without additional time. A Quora answer from week one still gets upvotes and clicks. Those hours are already getting cheaper in retrospect. An hour spent on paid ads does not do that.
The test I run on myself every week: is the ratio of output to input improving? In week one, 3 hours of work produced maybe 5 visitors. Last week, the same 3 hours produced the same new content but the old content added another 15-20 visitors on its own. The curve is bending. Slowly, but it is bending.
The founders who never see that curve are the ones who quit at week 4 because the current number looks bad. The ones who win are the ones who track the trajectory, not the snapshot.

30+ hours a week on a "side project" is basically a second job, and most people never quantify it that way. The fact that you tracked it puts you in a small minority.
The salary divided by customer count framing is brutal but honest. At my stage with roughly 150-200 sessions and no paying customers yet, that number is literally infinite. But I think the framing only works as a snapshot, not as a prediction.
The question I keep asking myself is whether those 30 hours a week are building something with a declining cost curve or just maintaining a flatline. If every hour produces content that keeps working after I stop touching it, the denominator grows while the numerator stays the same. If every hour is manual outreach that dies when I stop, the math never improves.
At what point did the unit economics start looking less ugly for you? Curious if there was a specific moment where the numbers flipped or if it was more gradual.

The publicly browsable dashboard is a perfect example of leverage work. You built it once and now every earnings season it works for you automatically. That is the exact pattern.
The finance Twitter result is painfully relatable. Two signups from hours of manual outreach is the kind of data point that forces you to rethink everything. But most founders never track it that specifically, so they keep doing the thing that feels productive without knowing it produces almost nothing.
The "feels productive vs actually compounds" distinction is the whole game. I have started asking myself at the end of each day: if I disappeared for a month, would anything I did today still be generating results? If the answer is yes, it was leverage work. If the answer is no, it was maintenance. Both are necessary but the ratio between them determines whether the business compounds or stays linear.
Curious how the dashboard traffic looks seasonally. Do you see spikes during earnings and then it drops, or does each season build on the previous one with more pages indexed?

That reframe from acquisition cost to research cost is spot on. Every manual interaction at this stage is data collection that paid acquisition can never give you.
I'm finding the same thing. The conversations I have on Quora and Reddit are not just driving traffic. They are telling me exactly what language my target audience uses, what problems they describe first, and what objections come up before they click. That is worth more than any keyword research tool because it comes directly from the people I am trying to reach.
The systematize part is what I am building toward now. Right now every answer is manual. But the patterns are becoming clear enough that I can start turning the most common questions into permanent content on the site. One manual answer becomes one article that answers the same question for hundreds of people without my time. That is where the CAC curve bends.
Curious what was the first thing you systematized that made the biggest difference?

$600+ per user when you count the hours is painful to see on paper but the fact that you tracked it puts you ahead of most founders who never run that math at all.
The leverage vs 1:1 split is a framework I wish I had from day one. I was treating all my time as equal until I started noticing that some hours produce results for one day and other hours produce results for months. A Quora answer I wrote three weeks ago still brings visitors today. A manual outreach email I sent the same day produced one conversation and nothing after.
Now I try to spend at least 70% of my time on leverage work. Content, site improvements, systems that keep running without me. The 1:1 work still happens but I think of it as the research budget that funds better leverage work later.
The docs and templates point is interesting. Did you find that those directly reduced your onboarding time per user, or was the bigger benefit that they attracted users who were more self-sufficient from the start?

"You're investing, not spending" is the right reframe. And the distinction between a CAC problem and a volume problem is something most people miss completely.
That's exactly where I am right now. The per-customer cost looks terrible at 156 sessions. But the infrastructure is already built. The calculator works, the content exists, the affiliate pages are live. Adding the next 1,000 visitors doesn't cost me proportionally more. It costs roughly the same daily time investment I'm already making.
The math that actually matters at this stage isn't cost per customer today. It's whether the cost per customer is trending down month over month. If it is, the model works. If it's flat, something is broken. I think most people quit before they have enough data points to see which direction the line is going.

That compounding check is a better framework than anything I've seen in a spreadsheet. "Is this hour making next month's hour more leveraged?" is the exact question that separates building an asset from just staying busy.
I'm in a similar spot. If I calculated my true hourly rate on this project right now, the number would be embarrassing. But every Quora answer I wrote last week is still bringing visitors today without any additional time from me. That doesn't happen at a day job. Your hour stays your hour.
The moment I stopped thinking about this month's ROI and started thinking about whether the system is compounding, the whole thing got less stressful and more strategic. Some weeks nothing moves. Other weeks one piece of content catches and suddenly the curve bends.
Curious how long you've been running yours and whether you've hit that point where the leverage becomes obvious?

That's exactly the distinction I keep coming back to. The question isn't "what's my CAC today" but "is the work I'm doing today making tomorrow's CAC lower."
Three weeks into my project, the honest answer is that my CAC is technically infinite because I haven't converted a paying customer yet. But the content I'm writing today is starting to rank, and the referral traffic from Quora and Medium already shows a 73% engagement rate. That tells me the compounding is starting even if the revenue hasn't caught up yet.
The inflection point you're describing is real and it's the part most people quit before reaching. Curious how long it took for you to hit that point with your automation systems?