Startups are about growth, and of all the different possible metrics, startups often focus on user growth - if people aren’t using your new product or service in greater and greater numbers, it’s a good sign that you’re not on the right track. And, as long as your business-model is solid, user growth is a leading indicator of revenue and profit growth.
But there’s another key metric that can be the making of your company, or the death of it - churn, or the proportion of existing customers who become ex-customers during a period of time (generally a year).
In the early days, it’s very hard to measure your churn because you simply haven’t been around long enough. The early revenue growth curves for high- and low-churn companies may look identical.
In fact, high-churn businesses often have exceptionally impressive growth numbers in their first months or years because they’ve found some viral loop, the crack cocaine of startup-land, which provides an explosive catalyst for new-user acquisition.
However, problems often start to appear after the first year or so - occasional weak months look exceptionally bad, since your business model relies on new customers in the top of the funnel to replace those flooding out. And even if you manage to keep finding new people to push into the top of your funnel for several years, you will eventually churn through your entire addressable market.
Then your business implodes.
The prime examples are Zynga and Groupon - both of which have been catastrophically failing over the last year or two.
When talking to startups, my favourite graph to look at is their revenue by monthly cohort. That is, splitting each months’ revenue by the month in which the users originally signed up to the service.
In a low-churn business, the graph looks like a layer-cake, with each new month of users adding a layer of revenue on the top of the business.
The thin cohorts at the bottom of the graph are the loyal users who signed up at the start, and who are still paying for your service 18 months later. The orange tip in the top-right are the users who just signed up, and have only contributed to revenue in month 18.
In the best low-churn businesses, the revenue provided by each cohort can actually grow over time - each cohort gets “fatter” - if you can successfully up-sell new services.
In a high-churn business, the cohorts decay alarmingly quickly, and once new-user acquisition starts to slow, the business implodes.
If you simply looked at the top-line of the graph (total monthly revenue), without considering the cohorts, you’d think the high-churn business was doing much better until at least month 13-14. But one-and-a-half years in, it’s clear that the low-churn business is on a path to success, while the high-churn business needs to execute a substantial pivot to avoid failure.
Indicators of Churn
So what are the characteristics of high- and low-churn businesses, and what techniques can you use to minimise churn?
High Switching Costs - if it’s slow, painful & costly to switch providers, you will experience very low customer churn. This is clearly the case with retail banks, email providers and recurring payment processors. I still bank with the same provider I chose 15 years ago, despite their crappy service and barely-functional website.
New competitors have developed clever tactics to solve this problem - developing onboarding tools to reduce switching costs, or focussing on new customers before they’ve committed to an incumbent provider. This is one reason Stripe focusses so hard on serving other startups, and banks target teenagers.
Sales Architecture - the way a product is priced & sold is often designed to minimise churn. Gyms & dating sites are the best examples here. People come out of the Christmas & New Year vacation period feeling unhealthy, lonely, and determined to change. One large dating site books over 70% of its annual revenue in the first quarter of the year.
Strong Brand Loyalty / Personality - the best brands position themselves as expressions of their customers’ personality.
Cigarette companies and beverage manufacturers are the best example of this - products are often indistinguishable from each other, and the switching costs are minimal. The customer can just reach for a different can on the shelf. Companies combat this by investing heavily in marketing to position their products in customers’ mind as extensions of themselves.
There’s an interesting phenomenon in the drinks industry where a couple of faceless mega-corps control a wide variety of incredibly powerful brands with basically indistinguishable underlying products - MillerCoors and Anheuser Busch.
Make Something People Love - after some fantastic comments from readers, I’ve added this extra point. Perhaps the most difficult, but also most effective form of customer retention is making a product or service that is dramatically superior to the competition. As Barry put it, “The switching cost is missing out on a great service.” It’s subtly different from brand loyalty - Google has very low churn because its search engine is so much better than the competition. It arguably doesn’t have a very strong brand or “personality”.
To some extent, you’re stuck with the churn inherent in your industry. But you do have some level of control.
If you enter a market like payment processing, be prepared for the long-haul. You’ll likely struggle to establish your credibility early-on, and finding innovative ways to acquire new customers be vital. You’ll need to raise significant amounts of early capital to invest in infrastructure that will only start to pay off years later.
If you’re part of an industry in which you can acquire customers incredibly cheaply and monetise them quickly, you’ll generally need less early capital. You’ll often reach profitability very early on, and it may feel like you’re on a startup rocket-ship. But make sure you’re providing sustainable, long-term value to your customers, otherwise you risk running out of fuel and crashing painfully back to earth.
Update: I’ve got a script that will auto-generate these charts from your Stripe.com data. If you’d like to see your own churn graph, email me at firstname.lastname@example.org for details. I’ll release the tool publicly once I’ve had a chance to test it some more.