The month my biggest customer left and took a third of my MRR

Every founder has one number they check before anything else. Mine was monthly recurring revenue, and in the spring of 2026 it read $24,100. The best it had ever been. What the headline number did not show was that one customer, a single mid-market account I will call the anchor, was paying $7,900 of it every month. Just under a third of the company sat on one invoice.

Then the anchor got acquired, the new parent already had an in-house tool, and I had thirty days.

This is a first-person diary of the month my largest customer left and walked out with almost a third of my MRR: the cashflow math I did on a napkin, the spending I had to unwind, and the concentration rule I now refuse to break.

Editor's note: this is a composite operator diary. The arc and the lesson are real and drawn from my own dashboards, but the figures are rounded and identifying details are changed. Treat it as one founder's honest account, not a benchmark.

Quick answer (2026): Customer concentration risk is when one customer, or a few, make up so much of your revenue that losing them undoes months of growth overnight. Common 2025 guidance treats a single customer above 20 percent of revenue as elevated risk. In my case one account was 33 percent of a $24,100 MRR, so when it churned my MRR fell to $16,200 in a single month, a 33 percent drop. What saved the business was not a growth tactic; it was being profitable enough on the remaining customers to survive the hit, then rebuilding revenue across more accounts so no single logo could ever do that again.

The chart I loved was lying about how safe I was

For most of the year the growth felt earned. I had gone from around $14,000 to $24,100 MRR in about ten months, and I told myself it was product-market fit finally compounding.

It was partly that. It was also one very good sales conversation eighteen months earlier. The anchor account had started small and expanded three times, and by 2026 it looked like this:

My top three customers were 55 percent of the company. I knew that number. I had even joked about it. What I had never done was ask the only question that mattered: what happens to my life if the top row disappears.

Trade-credit insurers and lenders have a rough rule of thumb for exactly this. Allianz Trade, writing for people who underwrite business risk, treats a single customer worth 20 percent or more of revenue as high concentration. M&A advisers are blunter: several noted in 2025 that above 30 percent, many buyers discount the business or walk. I was at 33 percent and calling it traction.

The thirty-day notice

The email was polite and short. The anchor had been acquired; the acquirer standardized on their own internal tooling; my contract would not renew at the end of the month. No anger, no failure on the product. Just a corporate decision two levels above anyone I had ever spoken to.

That is the quiet cruelty of concentration risk. My best customer did not leave because I did something wrong. They left because I had made my revenue depend on a decision I had no vote in.

Here is the napkin I did that night.

• MRR after the anchor churns: $16,200.
• My fixed monthly costs, including a part-time support contractor I had hired three months earlier on the strength of that growth: about $12,400.
• Net before: roughly $4,000/mo of breathing room. Net after: a few hundred dollars, and negative in any month with a surprise.

I had done the classic thing. I had seen the top-line climb and let my costs climb with it, most of that climb underwritten by a customer who was one board meeting away from gone.

Was I still default alive?

The framing that got me through the week was not a SaaS metric. It was Paul Graham's 2015 question about whether a startup is default alive or default dead: on your current costs and revenue, do you make it to profitability on the money you have, without needing anything new to go right?

Before the churn I was comfortably default alive. After it, I was on the line. The gap between those two states was almost entirely the spending I had added against the anchor's money.

So I did the unglamorous work first. I paused the contractor's hours for two months by agreement, moved two paid tools to their free tiers, and delayed a planned full-time hire I had already, stupidly, started interviewing for. Within a week my fixed costs were back under $9,500, which put me back to genuinely profitable on $16,200. I was default alive again. Everything after that was recovery, not survival, and recovery is a much calmer place to think.

This is the part the valuation articles never tell you, because they are written for the buyer, not the operator. Concentration does not kill you on the day the customer leaves. It kills you if you have already spent the money.

The month itself: what I actually did

Weeks blur, but the shape was this.

Week one: I did the cost triage above and told my two next-largest customers, honestly, that I was investing more in reliability and would love a call. Not a panic sale. Just closing the distance with the accounts I could not afford to lose next.

Week two: I read my own retention data properly for the first time since the month my churn quietly doubled the previous quarter. David Skok's line in SaaS Metrics 2.0 had stuck with me: get churn and customer happiness right first, or you are just filling a leaky bucket. I had been so focused on the size of the anchor that I had ignored three small accounts drifting toward cancellation. I saved two of them with a fifteen-minute call each.

Week three: I rebuilt the pricing page around more, smaller commitments. I added a mid tier that a self-serve customer could adopt without a sales call, because the whole problem was that my growth had required a sales call and therefore a whale.

Week four: the anchor's last invoice cleared. MRR read $16,200. I took the afternoon off, which felt irresponsible and was the best decision of the month.

The concentration cap I live by now

I did get back. It took about five months to return to $24,000 MRR, and the difference is entirely in the shape:

Same top-line number. A completely different amount of sleep.

The rule I follow now is simple and I enforce it even when it costs me a deal:

1. No single customer above 15 percent of MRR. When an account approaches it, I stop trying to expand that one and put sales effort into new logos instead. Slower growth, but the chart cannot fall over.
2. Big accounts go on annual contracts with a real notice period. If the anchor had been on an annual term with 60-day notice, I would have had a quarter to react, not a month. An exit should be a slope, not a cliff.
3. Spend against the floor, not the ceiling. I size my fixed costs to the MRR I would still have if my top two customers left. Everything above that funds one-off bets, not recurring commitments.

None of this is new advice. Every article on customer concentration risk says diversify. What none of them tell you is that the advice is useless in the exact moment you need it, because you cannot diversify in thirty days. The only thing that helps mid-crisis is the margin you protected before it started.

If you want the other end of this story, the founders who did not have that margin, it is in why we killed our SaaS at $12K MRR. The difference between their ending and mine was not the size of the shock. It was how much of the revenue we had already promised away.

If you take one number from this

Not your growth rate. This one: the percentage of your MRR that sits with your single biggest customer. Look it up right now. If it is above 20 percent, do not celebrate the size of that account. Ask the default-alive question instead, before your next hire and not after their notice email: without them, am I still standing?

I run a boring, diversified, slightly slower-growing company now. It is the most secure I have ever felt, and it started with one customer I did not get to keep.