A SaaS hit $50,347 MRR in March 2023. By November 2023, it was dead. The founder published a clean post-mortem on Medium and most readers walked away with the wrong takeaway.

The popular reading: "They couldn't compete with Datadog at the enterprise tier." That's true. It's also not what killed them.

I ran the post-mortem through the same churn analyzer I built for indie SaaS founders. Score: 18 out of 100. Grade: F. Three months of warning. Here's the honest breakdown.

The math, fast

DevOps Dashboard (anonymized in the original post) was a unified monitoring play. Datadog plus PagerDuty plus StatusPage rolled into one for backend teams of 5 to 20 engineers.

Peak MRR: $50,347
Active customers at peak: 2,400
Monthly churn: 8.2 percent (100 percent annual)
MRR loss per month from churn: $4,128
Burn vs revenue: $85K spent, $50K earned
Time from peak to shutdown: 6 months
Total lifespan: 15 months

The founder's own quote: "Every customer we acquired in 2022? Gone by 2023."

Why an 8.2 percent monthly churn rate is the actual headline

Industry benchmark for SMB SaaS sits around 5 percent monthly. They were 3 points high. Most founders read that gap and think "a little above average." Wrong frame.

8.2 percent monthly means your entire customer base resets every 12 months. You are running on a treadmill that speeds up every quarter. Acquisition has to outrun churn forever, and the moment growth slows, the math eats you alive in a single quarter.

That's exactly what happened. Acquisition dropped 60 percent. Churn stayed at 8.2 percent. The math finished them in 6 months.

What the cancel emails were saying

The post-mortem mentions support tickets jumped from 340 a month to 620 a month after the founder cut staff. That is the signal. Tickets up 82 percent with zero new headcount means every ticket got a worse answer, slower. Every bad support interaction is a future cancellation.

They had the data. They had the cancellation reasons sitting in their inbox. They didn't have anyone reading them. And then they cut the only people whose job was to read them.

Applying a churn health score retroactively

If they'd run their cancellation feedback through a churn analyzer at month 3, here's the score they would've seen.

Score: 18 out of 100. Grade: F.

Math (each insight deducts based on severity):

Critical: cut customer success entirely. Minus 20.
Critical: support tickets nearly doubled with no staff to handle them. Minus 20.
High: 8 consecutive weeks shipped zero features during decline. Minus 12.
High: ICP saturation, no path to next segment. Minus 12.
High: burn outpaced revenue 1.7x. Minus 12.
Medium: pricing locked at SMB tier when product needed enterprise floor. Minus 6.

Anything below 35 is F. This was a textbook F.

The 3 things RetentionCheck would've flagged

1. Cutting customer success during 8 percent churn is a critical insight

When you cut CS in a SaaS doing 8 percent monthly churn, you don't save $20K a month. You accelerate to 12 percent monthly churn within 60 days. The math gets worse every month after that.

The cancel feedback after the cut would've been full of phrases like "ticket sat for 3 days," "no one responded," "had to figure it out myself." Those are critical severity. Critical drops the score by 20 each.

2. A feature stall is a trust event

8 weeks with no shipped features during decline isn't slow. It is a death signal users can read. Customers compare your changelog to your competitor's monthly. Once yours goes silent, trust erodes faster than the product itself can break.

An exit survey would've caught this exact pattern. Question 4 should always be "what alternative are you considering." When 60 percent of cancellations name the same competitor, that's a positioning problem. When they name three different competitors, that's a relevance problem. Different fix.

3. ICP saturation is strategic, not retention

Saturating SMB DevOps teams (5 to 20 engineers) at 2,400 customers is real. But the founder's mistake was treating saturation like a retention problem. It wasn't. Saturation means new acquisition slows. Retention is the thing that keeps the existing $50K alive while you figure out the next ICP.

They tried to fix the wrong problem. Cut CS to extend runway, hoping to ship enterprise features. That move killed retention faster than the missing enterprise features could've saved acquisition. They optimized the wrong axis.

The pattern most indie founders miss

Cancellations don't tell you why customers left. They tell you what customers wanted you to do months ago.

A cancel email saying "didn't end up using it" isn't the cause. The cause was 6 weeks earlier when the user hit a friction wall and didn't get a response from your team.

DevOps Dashboard had this signal in their support inbox the whole time. 280 extra tickets a month carrying the exact reasons. They didn't read them. They cut the people whose job was to read them.

What you do with this

If you're running a SaaS at $5K to $50K MRR right now, three checks tonight:

Pull the last 30 cancellation messages. Read them in one sitting. Not skim. Read.
Pull the last 30 support tickets that came from users who later canceled. Look at the response time and the resolution. That is your churn root cause, not the cancel reason field.
If support volume is rising and headcount isn't, you're already in the failure pattern. Don't cut CS. Cut features instead.

Better: paste your cancellation messages into a churn analyzer. Get a Churn Health Score. See which insights are critical vs noise. Act on the top 2.

DevOps Dashboard had 6 months between the warning signs and the shutdown. That's enough time to fix it. They just needed someone or something to read the signal out loud.

Run this on your own feedback

Paste your last 30 cancellation messages at retentioncheck.com/try. 30 seconds. No signup. You'll see the same severity ranking, confidence scores, and direct quote backing that catches the pattern DevOps Dashboard missed.

Source post-mortem: "Our SaaS Hit $50K MRR Then Died in 6 Months" on Medium. Company anonymized in the original post.

How a $50K MRR SaaS died in 6 months (and what its churn data actually said)