The month support tickets ate my week at $25K MRR: a founder diary (2026)

By the numbers, that month at $25,000 MRR was my best ever. By the way it felt, it was the month the job quietly stopped being building and started being answering.

Quick answer (2026): At about $25K MRR and roughly 520 paying customers, I watched weekly support volume climb from around 25 tickets to around 46, which ate about 31 hours of a single week and left almost no time to build. The fix was not "hire someone" or "reply faster." It was a two-week support audit that showed three categories caused about 62% of every ticket. Fixing one bug, writing six help articles, and adding a self-serve billing page cut weekly tickets from about 46 to about 22. Support is a product metric, not a chore.

I run a small B2B SaaS on my own. A few weeks earlier I had let my one hire go at $22K MRR, a decision I still think was right and still wrote about honestly in the month I had to let my first hire go. What I underweighted at the time was simple: she had quietly been absorbing the support inbox. When she left, every ticket came back to me, and the timing was cruel, because the product was growing.

The week it actually happened

I keep a rough log of where my hours go. Most weeks it is lopsided but livable. This particular week, after a small release confused a batch of users, it broke.

Four hours of building. In a 44 hour week. And those four hours were not one clean block, they were the crumbs between replies, which is worse, because a support inbox does not just take time, it takes your attention in a way that never fully gives it back. Every ticket is a context switch, and a context switch is not free. You do not lose the ten minutes it takes to reply. You lose the ten minutes plus the twenty it takes to find your place in the thing you were actually doing.

By Thursday I had that specific, hollow feeling of being extremely busy and having produced nothing. I had felt the edge of this once before, in the stretch I described in burnout at $15K MRR. The difference this time was that the numbers looked great, so I could not even point at a fire. The dashboard said grow. The inbox said drown.

The thing nobody tells you about growth

Here is the part that took me embarrassingly long to see. Growth does not arrive as a bigger revenue number. It arrives, first, as more tickets. Every new cohort of customers brings the same confusions as the last one, and if you have not fixed the source, you just answer them again, by hand, forever.

And the customers who write to you are not even the majority. In a widely cited 2017 Harvard Business Review piece, the authors found that 81% of customers try to handle things themselves before contacting a live person. Help Scout's 2025 roundup puts the same instinct at 69% of consumers who first try to resolve issues on their own, while noting that fewer than a third of companies actually offer decent self-service.

Read that the way an operator should. Most people who could ask, do not. They poke around, fail to find the answer, and only then open a ticket. So a ticket is rarely a person who wanted to chat. It is usually a place where my product failed to answer a question it should have answered on its own. That reframing changed everything, because it turned my inbox from a workload into a bug backlog.

The support audit

I stole the method from an old rule of thumb that keeps circulating in founder communities: before you try to reduce tickets, spend two weeks tagging every single one. No fixing yet. Just labeling. You cannot cut what you have not counted.

So for two weeks I tagged all of it. About 92 tickets. Then I sorted the pile.

The top three categories were 62% of everything. Not vague, not spread thin. Sixty-two percent of my week was three problems wearing forty-six different faces.

The CSV one stung the most. A single export bug, triggered by a specific character in a customer's data, had been quietly generating about one in eight of all my tickets. I had been hand-answering the same corrupted-file complaint for weeks, treating a bug as if it were customer service.

What I did with it

Once the pile was sorted, the work stopped being emotional and became a to-do list. Three moves, in order of leverage.

1. I fixed the bug. The CSV export fix took one focused afternoon. It permanently removed about 12% of my incoming tickets. One afternoon of building erased weeks of answering. That is the whole argument for the audit in a single line.

2. I built the self-service the customers were already looking for. I wrote six short help articles for the exact setup questions in the top category, and I added plain empty-state hints inside the product at the two screens where people got stuck. This is the boring, unglamorous move that the HBR and Help Scout numbers above quietly beg for: if most people try to self-serve first, give them something to find.

3. I made billing self-serve. I added a simple account page where customers could download their own invoices, add a VAT number, and change seats without emailing me. The entire billing category, gone from my plate.

What changed

The next month, weekly tickets fell from about 46 to about 22. I got roughly 14 hours a week back. Not to relax, I am not going to pretend that, but to build, which is the only thing that actually moves a small SaaS.

And the honest, slightly uncomfortable footnote: MRR kept climbing, to about $27K. I will not claim the help center caused that, because I cannot prove it. But my first-week setup completion rate measurably improved once the confusing screens got clearer, and activation is the least mysterious lever in this whole business. Fewer people got stuck. Fewer stuck people churned. It rhymes.

The takeaway

If I could send one thing back to myself on that Thursday, it would be this. Stop measuring support as a feeling and start measuring it as a product metric.

1. Track tickets per 100 active customers, not raw ticket count. Raw count always rises with growth and tells you nothing. The ratio tells you whether your product is getting clearer or muddier as it scales.
2. Once a quarter, run the two-week tag-everything audit. It is tedious and it is the highest-return tedious thing I do.
3. Treat your top three ticket categories as a bug backlog with a deadline, not as a customer-service workload to be endured. A ticket you can delete at the source is worth a hundred you answer well.

Hiring support would have made me faster at answering the wrong things. The audit made a lot of the questions disappear. At $25K MRR, solo, that difference was the difference between a business I could keep running and one that was quietly running me.

This is a composite founder diary. The company, segment, and exact dates are anonymized, and every figure is self-reported and rounded to protect the founder's privacy. The pattern, the arithmetic, and the lesson are faithful to real solo-SaaS accounts, including my own. Nothing here is a real named person's verbatim quote.