on a journey to PMF: i stopped trying to explain it.
some months i can't draw a line through the data. i'm starting to think that is the line, so how do i embrace this.
Hey everyone,
Took a few extra days between issues this week because we were down in Austin for my nephew's first birthday.
Last issue I pulled a quote from Alex Turnbull at Groove. He called his company an “NRR zombie” — $4.5M ARR, solid retention, healthy NPS, and MRR down every single month for 14 months.
His line that stuck with me: the market is slowly moving while you keep the lights on.
I’ve been sitting with that frame for a couple weeks now, and I feel like his description of Groove is a very strong description of Journey.
Our best customers love us. They add users every month. They use the product for the reasons that generate the most value, but I’m getting fewer and fewer of these customers. More and more of the customers that aren’t our best… they churn faster.
The market is moving (I think faster than I would like) while I’m trying to keep the lights on.
There is something to be done with this. I’m not exactly sure what, but I’m thinking through it.
in an AI world, SaaS without PMF makes less sense than ever before
In the last 30 days on Journey, I’ve had:
A handful of upgrades from customers I haven’t talked to in months
New paying signups from industries I wasn’t targeting
One churn from an account I was half-sure would become a reference story
I’ve tried to draw a line through these. I can’t. Just motion, in both directions, cancelling itself out most weeks.
the 50/50 answer i can’t give
Last week I was on a call with a prospective buyer kicking the tires on Journey. He asked a question I’ve been dodging in my own head for a year: why do some customers care about how their content is packaged, and some don’t?
I told him the truth. It’s about 50/50. Half my customers hear the pitch and say “I really, really, really care about how my stuff looks when I send it out.” The other half shrug and say “I don’t care, email works fine.”
I’ve never cracked the signal that tells me which half a person is in before I get on the call.
Firmographics don’t predict it. ICP filters don’t predict it. I’ve built signals models in Clay trying to. None of them hold up. People buy software for reasons they can’t fully explain to themselves — let alone to me.
what alex got right
Here’s where the zombie frame lands for me.
The surface looks stable. The net numbers say fine. But underneath, something is moving. The market is moving. The buyer’s emotional model for how they pick software is moving. And while all of that is moving, I’m keeping the lights on and calling it retention.
Staying the same is not staying the same. It’s a slow drift you don’t see on the dashboard until it’s too late to react.
The scary part isn’t the drift itself. It’s that the drift isn’t a logic problem I can out-think.
the lever i actually have
If I can’t reliably grow demand I can’t explain, the only variable fully in my control is how long I can stay in the game.
That’s burn rate.
I’m mid-audit, but the list I’m working through right now:
Growth experiments. Paused or killed the ones without a clear signal. No more “let’s see” budget.
Attio. Reconsidering whether we need the full CRM stack or if a lighter setup gets us 90% of the way there.
Customer.io. Same question — are we using enough of it to justify it, or can we consolidate into something simpler?
Polytomic. Data sync is real, but so is building the pipe we actually need in n8n for a fraction of the cost.
Raaft. Nice-to-have retention flow. Not load-bearing.
None of these are bad tools. They were the right call when we were spending like the market was still growing up and to the right. They’re not the right call when the signal is “stay in the game.”
The AI-native founders I’ve been paying attention to aren’t special because they found some trick. They refuse to rent leverage they can build themselves with Claude, n8n, and a few Claygents stitched together. They own the automation instead of subscribing to it.
That’s the operating model to manage the ambiguity and the lack of patterns that are more dominant in businesses without PMF in an AI-first world.
what i need to build on
Alex’s post had a line underneath all the hard numbers that stuck: your best customers still love you.
That’s my situation too. Power Knot is still paying. Long Angle is still embedding Journey into their whole community experience. Half my calls end with someone saying they really, really care.
I just can’t predict who the next one will be, so I have to cut the fat. Let the randomness run. Trust that the customers who care.
That’s it for this one. Let’s keep it going.
Cheers, Danny
fun stuff
We were in Austin last weekend celebrating my nephew’s first birthday.
The other thing that made this week: my son’s first school dance as a kindergartner.