A growth framework for early-stage consumer subscription startups (a WIP post)

Note: I’m putting this post together in chunks and publishing it as I finish each chunk. Think of it as shipping quickly.

As a founder, there are times when you sit down alone and think “what do we need to work on next to grow.” If you’re anything like me, 500 things come to mind at once and it feels overwhelming to try to find the signal in the noise.

As I’ve returned to this question countless times, poured over the “literature” from the best thought-leaders on the topic, and measured my own tests, I’ve synthesized a framework for growth that’s helpful for me (1).

The framework takes the shape of a decision tree. The idea is that you can take your company through the decision tree and get a pretty good directional idea of where to put your chips.

The first hinge question: “do you have product-market fit?”

One of the great things about a consumer subscription business is that you can measure product-market fit mathematically at a very early stage by grading your cohort churn rate (2). Depending on where your churn rate lands, you have two options:

=> if your monthly cohort churn rate is >5%, then you don’t have product-market fit (3). Your bucket is leaky, and your focus needs to be 100% on product-market fit and 0% on scale. This is the hard road. Depending on how far away from 5% you are, you may need to reevaluate your value proposition entirely. David Pakman from Venrock penned my favorite article about this topic, and he puts it pretty bluntly:

In my experience, the only way to meaningfully reduce churn in high-churn businesses is to essentially redesign the service and the value proposition. For clarity, I have looked at close to a hundred subscription services and have never seen churn improve from, say, 12% a month to 7% a month without fundamentally changing the service

I obviously don’t have the experience that Pakman has in evaluating subscriptions, but if I found myself in this scenario, my two cents would be to follow Rahul Vohra’s general advice: isolate the people who are loving your service, and only listen to what they want + acquire more of them until you get your churn under 5%. If you can’t get there, accept the math of your business and act accordingly.

=> if your cohort churn rate is <5%, then you need to put your chips on growth, which leads us to the second hinge question.

The second hinge question: “have you cracked paid?”

Why Paid?

Growth is going to break down into three critical vectors that you should solve for early in the growth process: 1. referral growth 2. paid growth 3. unique product flywheels. There will always be ancillary channels, but these are the three to focus on in the beginning.

Of those, I would argue that paid is going to be the most powerful to focus on. Here’s why: referral growth is a function of your core value proposition, market positioning, CTA frequency, reward, and friction to referral. In other words:

referral growth rate = (value prop + positioning + CTA frequency + reward )/friction

Of those factors, value prop and positioning are the most difficult to improve. But here’s the thing. If you have a sub-5% churn rate, your value prop is pretty good—good enough that current customers will be telling new potential customers somehow. And here’s the other thing. When you work on paid, you have to focus relentlessly on your market positioning because you’re trying to convert people with no knowledge of your company into subscribers.

That leaves CTA frequency, reward, and friction as the pressure points that accelerate referral growth. And while these require brainpower to solve, they will only incrementally accelerate referral growth, while paid will unlock a brand new and almost infinitely scalable channel (4).

Cracking Paid: Efficiency

So what does it mean to “crack” paid? I think it boils down to two things: 1. efficiency of paid growth and 2. velocity of paid growth.

By efficiency, I mean CAC and payback period. Here’s how I think about it.

This is a simplified model here of a consumer subscription business with a 50% gross margin and 5% churn — a business we’d say has P-M fit in the mathematical sense. Let’s also assume that they have a $10 monthly price, and $84 annual plan ($7/month), and that 25% of customers buy annual when they join. Last, let’s assume that they’re giving every new subscriber a free month trial.

There are two different ways of looking at payback here that illustrate what it means to “crack” paid.

If you look at your payback on an amortized basis (5), you can see that by month 6, you’ve returned roughly $20 to the business. So, if you’re paying roughly $20 to acquire a new paid customer, you achieve breakeven in 6 months and close to a 1x return inside of the first year.

But if you look at paid on a cash basis (6), you’re actually getting $20 inside of 4 months. This is the more important way to frame payback, because it illustrates how much of a cash machine you can create with paid growth. At a CAC of $20, you can reinvest that money after only 4 months to buy another subscriber, which creates a snowball effect over time.

Payback period (aka your paid efficiency) is going to dictate how fast you can go. If you’re waiting too long to get your cash back, you can’t invest quickly in new growth. If you’re getting your cash back immediately, you’re creating a cash machine that snowballs.

This math also illustrates why having a well-priced annual plan is so critical. When you drop the price of annual, you increase the percentage of members who purchase annual, which decreases your payback period and drives the efficiency of your cash machine:

Here, the example business dropped their annual plan to $6/m ($72 annual) and decreased their payback to 3 months (7).

Cracking Paid: Velocity

The second dimension to paid is your velocity of acquisition. If efficiency is your MPG, velocity is your MPH.

As I mentioned above, the reason that cracking paid is so critical is that it unlocks an additional channel. The implication is that you’re adding to your growth rate, which increases your maximum MRR/ARR.

Continuing with the example above, let’s say you “turn on” paid and begin acquiring 50 new subscribers/month at a monthly churn of 4% and an ARPU of $8.40. In this case, you’ve increased your theoretical max MRR by:

(50 new subs/month)/(4% churn)*($8.40 ARPU) = $10,500 MRR Increase

But we’re not dealing with theory here. As a founder, you care most about the MRR increase that occurs inside of the first year or two. A good rule of thumb is that you’ll boost your MRR by about 40% of your theoretical max within a year, and about 60% by the second year:

MRR increase over time (in months) for the example above

So, 40% of $10,500 is $4,200 and 60% is $6,300 and you can see that maps to the graph at 12 months and 24 months, respectively.

Back to our example then: for every 50 members/month you increase your growth rate, you’ll be increasing your MRR by ~$4k in the first year, ~$6k in the second, and so forth. This is why understanding the velocity of paid is so important. If you’re stuck at, say, $30k MRR and you want to level up to $60k MRR in the next two years, you really have to be able to ratchet up acquisition through paid:

($30,000 Target MRR Increase)/(60% Achieved in 2 Years) =>($50,000 Theoretical Max Increase) / ($8.40 ARPU) * (4% Churn) => (+238 members/month growth rate)

Continuing with the example, if you are spending $20/member, you’d be increasing your marketing spend by

($20 CAC) * (238 Members/Month Target) = $4,760/Month

In order to achieve an MRR increase of $30k/month over the next two years. Not a bad investment (8).

To bring it back around, cracking paid from a velocity perspective means that you’re able to invest a high enough volume of marketing dollars to achieve a step-change in MRR (e.g. 50+% increase in 2 years). When you combine this with cracking efficiency, not only are you leveling up your MRR, but you’re doing it in a way that creates free cash flow to invest. That’s a lethal 1–2 punch.

Beyond Paid

Let’s go back to our decision tree. If you have mathematical product-market fit and you’ve cracked paid (velocity and efficiency), then you’re in a very rare cohort of startups and you’re beyond the scope of this type of post. From here I can only offer conjecture since Offline isn’t quite over the hump yet.

  • Unique product flywheels
  • Organic

5.5.2021 the sidewalk ends here. I’ll be back with more.


(1) As in, for consumer subscription startups. Extrapolate at your own risk.

(2) If you don’t know how to grade your cohort churn, you need to back up and do some reading + install an analytics package that exposes the necessary data. Helpful: https://mixpanel.com/topics/cohort-analysis/

(3) It’s worth pointing out that both Netflix and Spotify had churn rates that were 10%+ in the early days, so the 5% rule and Pakman’s accompanying thesis are not foolproof and there are exceptions. Here are the two exceptions that I think are valid. 1: you’re growing virally. If you have a freemium service like Spotify that is growing rapidly, you can afford to take more risks with product-market fit because you’re building an enormous funnel of free users to continually test your value proposition on. 2: you’re spending an enormous amount on customer acquisition, and your most engaged users are demonstrating a much lower churn rate. In this case, your userbase is bifurcated: people who “get it” are sticking at an impressive rate, and people who are “sampling” are churning at a high rate. This drags your average way down, but peeling back the layers reveals a more compelling story. Jason Calicanis talks a bit about this here.

(4) Obviously, paid is not a panacea and it gets less efficient as you scale and saturate the market. But those are down-the-road problems that 99% of startups never even get to experience because they aren’t big enough yet.

(5) Meaning that you take that annual payment and consider it as “deferred revenue” that hasn’t actually been paid back yet and instead gets peanut-buttered across the year.

(6) Meaning you recognize the full annual payment immediately. In this case 1 in 4 members are buying annual at $84, so they’re contributing an average of $84/4 = $21 during month one. Looking at the world this way can be a little dangerous from an accounting perspective because members can ask for an annual refund, but we’ve already assumed at this point that you’ve got P-M fit, so your refund requests should be very low for annual members.

(7) Obviously dropping your annual price can cannibalize your LTV. You have to decide what you’re optimizing for. You also need some time to measure the churn of annual members vs monthly. With P-M fit, don’t be surprised if annual members have a lower adjusted churn than monthly.

(8) Of course, in order to maintain 150 members/month and a $20 CAC, you’ll be rotating ad creative, working on CRO, etc., which costs more money. But, as I pointed out above, the effects of these costs are not isolated to your paid channel. Investment in market positioning and conversion will have ripple effects across the business that will make your overall marketing machine more efficient.

Founder & CEO at Offline