On Churn

2026.07.12

CLX

[No Investment Banker Dinners; Auto Pilot Acquisitions; PMF is a Gradient; A Churn Wall; Too Much Value?; NRR; Lead Bullets; Silver Bullet]

Thesis: Product market fit is a gradient that you need to keep iteratively working toward, and solving churn is a big part of it.

[No Investment Banker Dinners]

WhiteWhale is coming up on it's 2 year anniversary.

And really, the underlying problem technical problem is something I've been involved in directly since 2022, and more indirectly for a good chunk of my life.

Given all that, sometimes I think we’re a bit delusional for continuing on the quest even though we don’t have Product Market Fit (PMF) yet.

For those unfamiliar, Marc Andreessen describes PMF as:

The great news is we’re certainly NOT strictly in Andreessen’s first bucket. But while we have elements of it, we’re not fully in his second bucket, yet, either, though.

This makes sense: PMF is a gradient, not a binary thing. And you get more mixed signals than in dating. A challenge is sifting through the information and finding where you land on the scale.

If you think you’re further from PMF than you are, you might over correct and pivot and fuck up the thing that was working. If you think you’re closer than you are, you might lever up to fast and get top heavy and crumble!

The good news for WhiteWhale we're closer to PMF than every other time I've written about it. We have more positive feedback and word of mouth, we’re getting intros to decision makers in our space, we're hitting all time high revenues, we're collecting customers faster than ever, and we can acquire some customers without even having to get on a sales call!

Still, we're not there. Churn is too high and usage is not expanding enough. And, the investment bankers aren’t buying us free meals, either.

So, we're going to do what we always do - tackle our problems directly and aggressively while looking for a more elegant solution that takes us a step function closer to what we want.

[Auto Pilot Acquisitions]

First things first, I've been yapping about PLG over the last 3 months now.

The really good news is that it's still actually working! My note a few weeks ago ended up not being a false alarm. There are still a lot of hiccups and optimizations, but there are some people who are buying the product and correctly provisioning it (and to be fair, we've also rewritten the product to get it a lot closer to correctly provisioning itself!).

There's still a lot of progress to be made, like getting rid of some edge case bugs that can be quite bad, or introducing a support bot, but the infrastructure is there. That, plus now having a completely automatic affiliate program makes our lives a lot easier. There are still onboarding calls for some people and Jack is still fielding plenty of old deals or prospects from his dms, but we've seen enough to know that full self service is possible.

It's just a function of further refinement and iterations on our end.

This is absolutely lovely news, because it means we aren't inherently needed to get on sales or onboarding calls to scale the business. The product can work on it's own, and in an increasing number of cases, will.

This let’s us focus less time on linear activities like sales calls and onboarding, and more time on asymmetric investments, like product work & marketing & funnels.

[PMF is a Gradient]

You can have more or less PMF. We have astronomically more PMF now than we did 2 years ago, and more than we have at any point in our history.

The mixed signals can make it really hard to understand, though. As an example, last week, in a 48 hour period, we had:

• 3 customers churn

• One customer on an annual subscription tell us that they closed a deal directly attributable to us that covered the cost of the subscription

• Another customer notice a bug in the software but then ask us how he can help promote us out in the real world

• A text from a friend saying that someone in his co working space is raving about how awesome the product is

This is without mentioning the ups and downs from hot patching product issues on the spot, etc.

The point is, the positive stuff is really positive, but the negative stuff is quite negative, too!

And you may be saying that the churn is natural; of course customers eventually leave! But, if you're not careful, churn can be so high that it fundamentally blocks your ability to grow.

The question becomes how do you get MORE of that great feedback and much less of the churn?

[A Churn Wall]

You can calculate your maximum revenue based on Churn and revenue growth.

When I say Churn, I mean what percent of your old revenue you lose each month.

As a fictional example, if you have $60,000 in MRR in Month 1, and then $54,000 of that revenue stays til the next month, your churn is 10% (($60,000 - $54,000) / $60,000).

Note that churn doesn’t consider the new revenue you add in the next month. If you add $6,000 in new revenue, you’d end up at $60,000 in revenue at the end of month 2—but this is literally how much you started with! So, even after doing all of the effort of adding $6,000 in revenue, you’re running in place on a treadmill!

So, Churn & revenue added each month actually show you the ceiling of your business.

If you have 10% churn and $6,000 added in net new revenue each month, your maximum revenue is something $60,000 per month.

$60,000 per month, though, is really hard to hit with those numbers. Every month that you get closer, your rate of progress slows (at $45K / month, you lose $4.5K per month and add $1.5K to end up at $46.5K. But then, the next month at $46.5K, you lose $4.65K and only increase by $1.35K!).

The only way to overcome this ‘churn wall’ is to decrease churn, increase revenue added, or both.

This is why losing 3 customers in one day is something that seriously weighs against all that other good news I mentioned above. If we add even one new customer a day, then losing 3 at once puts you at nearly 10% churn for one month's cohort*!

And yes, total churn is calculated not just with one months cohort... but, all I'm saying is I'd be telling you our total churn if it was single digitals…

Showing a weakness like this publicly on the internet might seem crazy, but, I don't care for 3 reasons:

1. It'd actually be worse to share if I said I figured it out - competitors would be more inclined to copy me!

2. I don't think a lot of new SaaS companies are doing much better, especially the ‘ai’ powered ones

3. It's a problem we're spending all of our free working hours to solve. We will solve it.

_{*One way to help understand churn is to break users up into 'cohorts', or groups with something in common, like the month in which they signed up. Even for a rapidly changing product, these people will have a more similar experience than randomly grabbing 5 people from every month!}

[Too Much Value?]

Last week, I got on the phone with a new customer. He'd been burnt by a quasi competitor charging him 9 times what we are. This competitor kept saying he could do everything and continuously fell short; he ignored my Father's oft repeated advice "Under promise, over deliver."

Anyways, this client is quite technically apt & operates a lot of stuff out a central system he built with an AI in the middle. He said two very important things.

First: "I can't build this myself."

This is quite awesome coming from the kind of guy who is always asking what he can build and what he can't; it's good validation that the challenge of recreating our software is far more expensive than the $200 he's currently paying.

Second: "You're giving me too much value."

This is something we've always known and something my sales coach has warned against. It's the reason we started doing quarterlies and annuals at one point: if we're enriching and ranking an account list, we give a lot of value upfront!

Especially in his case; he had 36,000 records, and we showed him the best 1,400. Ranking 36,000 records would be insanely expensive for many of our competitors. Granted, we only let him see the in depth data on 300 of the accounts this month, but he still knows the names of the other 1100 & gets some useful info about them.*

Similarly, in the last 2 weeks, we heard from a rep at a client that used us for 3 months: he told us that after they got rid of us, the client still closed a deal off of our data, giving them insane ROI.

In some ways, this might be the best news we've gotten... some portion of our customers are churning because they're getting everything they need right away. Meaning not only is it valuable, but it’s too much value at once.

[NRR]

Net Revenue Retention (NRR) is another concept related to Churn; it's how much of last month's revenue you kept on a percentage basis. Meaning, if your churn is 10%, NRR is 90%.

However, NRR can be positive, too. If you have 10 customers each paying $200 per month, and you have one leave but 2 double their subscriptions, NRR becomes 110% (($2000 - $200 + $200 + $200) / $2000).

If you have positive NRR and do nothing, the business will just grow. Moreover, this means you can neutralize churn by up selling your best customers (the ones who want to brag about you at conferences) rather than just by reducing the number of customers who leave you.

Every customer who doubles thei subscription buys you back one who leaves.

I bring this up here, because if some customers think we're giving too much away at the price point, this represents a very exciting opportunity to restructure the offer and up sell them.

As of now, our product is structured horrendously poorly for up sells. So, this is something we're going to fix.

[Lead Bullets]

There are so many little wins we've been able to add in the last week or two that should help reduce churn:

• Fixing a lot of the deepest & most annoying & confusing bugs

• Automating a lot of the activation for people in the PLG flow

• Adding a welcome tour that takes them through the platform with videos of Jack yapping about how to use it

• Clearly setting the expectations around the limitations of the product during said onboarding flow

• Adding a clear help center to the settings

• Adding a couple emails that go out after you buy, including one talking about success stories

Most of this stuff falls in the camp of "obvious" fixes that we were clearly missing and not prioritizing enough.

From one point of view, doing a lot more automation of the activation should help a ton. Customers who never activate have a 100% chance of churning, which is the highest of any category; so, if you make sure every one activates, then of course you'll reduce churn!

From another angle, if you view churn as a build up to some critical mass of confusions and frustrations that causes a customer to eventually leave, this stuff adds up a lot, too. It gives them pressure release valves and reduces the probability of that critical mass being hit.

Here’s a good encapsulation of this grindy, solve the issue at all costs mindset:

Subscribe here for my weekly rant that ranges from startups to philosophy to chasing truth to living a good life and creating cool things.

[Silver Bullet]

Still, below all the lead bullets, you can't help but wonder if there is a silver bullet. Is there some step function that can solve a ton of the problem at once?

At any combination of product and offer, there is some minimum churn and maximum NRR you can achieve by optimizing and making the above grindy changes mentioned.

We think there is another combination of product and offer with a much lower possible churn and higher NRR, though. Something that will feel a lot more like Mr. Andreessen’s quote above… something that might bring the bankers out in droves.

So, while we're grinding away on the lead bullets, we're also testing out a silver bullet that involves changing the front end and the way that we price and deliver value.

The goal is to build something that makes it easy for all of our customers to get as excited as our best customers. And to sell it with an offer that makes the guys who says we're giving away too much value actually feel like a more reasonable trade while still being impossible to recreate. We're pretty thoroughly convinced that these two elements, the lack of the up sell and the churn, are actually both closely related and solvable with one big change to product and offer at the same time.

This is challenging, because as I've said before, most our closest competitors haven't seem to have discover a low churn / high NRR business model, either.

The trouble is, a lot of them used VC money to grow as big as they did without ever solving this problem. You can buy a lot of things in life, including customers. If you could buy PMF, though, VC as an asset class wouldn't be in such a weird spot.

So, in other words, if Jack & I took on money for WhiteWhale, we'd have more revenue, but we'd also be losing money and likely dealing with the same issues we are now. Just at higher stakes, because if we didn't solve it fast enough, we'd run out of cash.

With a profitable, bootstrapped business, this is not a risk. We have all the time in the world, but we're still acting like the sky is falling tomorrow.

So, even being nearly two years into a business that still doesn’t have PMF, I’m more excited than ever.

Live Deeply,