“We’ve got great repeat rates: over 60%” a founder told me recently.
“That’s great,” I replied “so 60% of your customers you acquire repurchase the following month?”
“Errr.”
There's a pause, then some questions, as we get to the crux of how that’s measured.
“60% of our revenue last month was from repeat customers” the founder eventually tells me.
“OK that’s a 🚩” I reply and the founder looks confused.
Unless you’re Unilever or Coca-Cola, a repeat rate like this is a sign you’ve not figured things out yet.
This is counterintuitive—especially for those with DTC scepticism – but high repeat rates like this are a real problem for your business.
And I think Shopify has lots to answer for with this issue.
Here’s why.
This intro might feel unfair to the founder in question. But the truth is, I’ve heard variants of this story for the last decade—and it's increased since Shopify became mainstream.
“We’ve not spent a penny on marketing so far”
The “we’ve got high repeat” statement is on par with “we haven't spent on marketing yet.
Both are told as things that seem good but are in fact bad. If you’ve not spent a penny on marketing so far, for almost all kinds of businesses that means you’ve acquired (1) network, (2) low hanging fruit, and (3) maybe some high intent cold outreach. But you’ve not got a scalable business model.
The same is true with this repeat rate issue.
Repeat disguises how good your acquisition is
We can all hide behind repeat. Maybe you’ve been around for a bloody long time and so you’ve naturally built up a great back book of revenue.
Maybe you’ve got other sales channels and so they’ve migrated for you over to DTC.
Maybe you’ve had some wins with PR or influencer or organic in the past and that’s created some back book of customers ready to buy again.
But what it doesn’t show you is how good your acquisition engine is.
Consider this chart: repeat as a % in this example is growing.
If you’re a business that’s tapped out its acquisition, that’s potentially a great thing. That means you’re making more money each month from that huge cohort of people who have bought from you already.
Now let’s look at that chart as revenue stacked together
Here we can see two trends:
Repeat as a % of revenue is increasing
New customer revenue is decreasing – and your repeat isn’t catching up to fill in the back book.
This is not a sustainable growth engine.
If I were an investor that repeat % is suddenly a big flag because it hides actually that this business hasn’t yet found a way to acquire customers.
But is retention needed to make a business work?
Yes it is.
One question is: what type of business do you want to become?
Let’s frame part of this question as ‘what do you need to hit £50m ARR?’ Now only a tiny percent of businesses will ever do that.
And the majority of consumer businesses out there would be far better not aiming in that direction. For lifestyle businesses, hitting a few million in turnover would radically change the lives of the owners.
But let’s say you’re on a growth hunt. £50m ARR is half way towards racking up a £1bn valuation.
If you make very high value items, you don’t need much repeat.
Sell a sofa or a mortgage or a luxury watch or a holiday, you can get away with selling just one thing to your customers each year. Maybe there will be some repeat every few years, but maybe not.
If you make £10k from your customers per year, then you don’t need to worry about retention. That said, if you do get that model right you’ve got a winner – note why Salesforce has a $266bn market cap because it nails both high value and high repeat.
Most consumer businesses which are being built today are attempting to occupy a land that sits between £100 and £1,000 of revenue per year per customer.
If you’re at £10 per year, you’re in trouble. Look at the chart above – notice what companies make £10/year work? Those with billions, hundreds of millions, or at the very least, 10s of millions of customers. You need 500,000 just to make £5m a year if your average revenue per user is this low.
So yes for the majority of brands, retention and repeat is important. Which is why it’s even more important to measure it right.
The two most important ways to measure retention
Measuring retention is too important to get wrong. And there’s two important ways to look at retention. One of which is almost impossible to find inside Shopify – hence why I think Shopify has been part of this analysis issue.
How to measure retention rate
Here’s what you need to measure:
The total number of first time purchasers you have
Of that cohort of people who went on to purchase a second time
Of that cohort who went on to purchase a third time
and so on
Get that information, and you might end up with a table or chart like this:
This is useful because it allows you to understand where in the journey people are dropping off.
Blackhat gaming 2nd purchase is easy: sign people up to a subscription and don’t notify them of the repeat. You’ll have massive 2nd purchase rates, but likely big drop off immediately after.
Typically speaking, you will expect to see repeat drop down each cycle as cohorts naturally churn over time.
The rate at which this happens determines how strong your retention is.
With this retention work, you can understand:
Where the biggest dropoffs are
What the true repeat rate is
If your second purchase repeat rate is, say, 15%, and you have a low AOV that requires repeat purchases, that's a red flag.
At Wine List, our retention figures based on the above were approximately:
Purchase 2: 55%
Purchase 3: 65%
Purchase 4: 75%
Purchase 5: 85%
That might seem good but consider this:
Let’s imagine that you take a £19 order every month and you acquire 1,000 customers in this business.
Your retention rate flattens which looks great but by the time it does you’re now only making £5 of revenue per original cohort member.
I.e. if you acquired 1k customers in the above model.
Purchase cycle 1 you make £19k
Cycle 2 you make £10.4k
But by cycle 5, you’re now only making £4.3k – you’ve lost 75% of the cohort value already.
Whether that works for your business or not comes down to your contribution margin and acquisition cost. If your CM is 60% and you’re acquiring customers at £11 then you break-even on cycle one and the rest is off the races.
But if your acquisition cost is £30 and your CM is 50% then it takes you 9 purchase cycles to break-even.
The X factor of time between purchase cycles
The above model fits really nicely if you’re a monthly subscription product. But lots of products aren’t. Lots of products don’t naturally follow subscription patterns and rely on customers coming back to the site.
Let’s imagine you run this business:
£20 CPA
60% CM2
Based on the graph you can see that you break even somewhere around the third purchase level. Happy days.
But how long do you have between purchase cycles? If it takes three months for people to use your product fully, now you’ve got the problem of cashflow. Your retention is great, but cash flow will determine how quickly you can grow. If it takes nine months, that’s going to require serious capital investment to get yourself towards your 500,000 customer mark.
Why cohorts become important
The above is important – but so too is looking at this on a cohort basis.
The people you acquire earlier will have a higher need state than those you acquire later.
That means you need to work harder to retain them. You need to focus on retention and actively try to improve it all the time.
A chart like this becomes useful
In this example we can see the retention rates changing over time.
In an example like the above, here are hypotheses I’d investigate:
Q4 hasn’t had the chance to repeat yet – indicating long purchase cycles / time between purchases
December was likely high gifting and so unlikely to follow usual behaviour
Retention rates were improving throughout 2024 – both at 2p and 3p – indicating active improvements to retention.
This allows you to model out your retention behaviour much better in the future.
What is your true revenue per customer?
A side bar to this whole conversation is what is your average revenue per user/customer (ARPU).
Look at your January cohort from last year and sum the entire revenue from that cohort, then divide it by the cohort size.
Too often I see ARPU calculated as “good customer behaviour is 4 purchases per year, therefore 4X AOV = the ARPU.”
Summary
% of revenue is a bad metric—it masks your acquisition performance, which should consistently be growing.
Retention rates are important to understand – but model them based on purchase cycle
Time between purchases dictates your cash flow needs for growth
ARPU = total revenue from a cohort / cohort size
P.S. If you enjoyed this, please tap like, leave a comment, or share this with a colleague.
P.P.S. We’re hosting an event tomorrow in SE1, this is your last chance to book a ticket.