Why your repeat customer rate is more valuable than your acquisition rate.

Walk into most growth meetings and you'll hear the same numbers celebrated. New customers this month. Cost per lead. Conversion rate on the latest campaign. The whole room is pointed at the front door.

Almost nobody asks the question that actually prices the business: of the customers we won a year ago, how many are still here?

I'm going to argue something stronger than "retention matters too". I'm going to argue that your repeat customer rate is worth more than your acquisition rate. Not equal. More. And I'll show you the actuarial reasoning behind it, because that's the lens I've spent over a decade applying inside banks, insurers and technology companies, and it's the lens most businesses have never pointed at their own customer base.

First, let's be honest about the famous statistic

You've probably seen the claim: a 5% improvement in retention increases profits by 25% to 95%. It's attributed to Bain & Company and it appears in roughly every retention article ever written.

I went looking for the source. The trail leads back to a 1990 paper by Frederick Reichheld and Earl Sasser called "Zero Defections: Quality Comes to Services". The analysis was built on service industries of that era, credit cards and insurance brokerage among them, and the 95% figure is the top end of a wide range across very different businesses. It's a 35-year-old finding that gets quoted today as if it were a law of physics.

I'm telling you this because I want you to trust the rest of this essay. I don't build strategy on borrowed statistics. The direction of that finding is right, and I've watched it play out in client data across hospitality, property, retail and education. But the reason it's right matters more than the number, and the reason is what most businesses miss.

Acquisition is a lottery ticket. Retention is a known hand.

Here is the core of the argument, and it has nothing to do with cost.

When you acquire a customer, you are buying a random draw from a distribution. You do not know if you've just won a customer who will transact for five years or one who will vanish after the first purchase. You paid the same to acquire both. Your acquisition dashboard counts them as identical wins.

When you retain a customer, you are not drawing from a distribution. You are holding a hand you can already see. You know their purchase history, their frequency, their margin, their trajectory. You know, with far more confidence than any lookalike audience can give you, whether this specific customer is above or below average value.

That asymmetry of information is the whole game.

Acquisition spend buys uncertainty. Retention spend protects certainty.

A rand spent keeping a customer you know is valuable will, on average, beat a rand spent acquiring a customer who might be.

Most businesses run this backwards. They spend aggressively on the lottery and passively on the known hand.

The actuarial angle: stop using one LTV for everyone

Actuaries have a tool for exactly this problem. Insurers value their book using embedded value: the present value of future profits expected from the policies they already hold. No insurer would value every policyholder at the average premium multiplied by the average duration. The entire discipline exists because averages lie. Risk and value are segmented, modelled and projected forward, per group, under explicit assumptions.

Now look at how most businesses calculate customer lifetime value. One number. Average revenue, times average margin, times average lifespan. The whole customer base flattened into a single fictional person who doesn't exist.

That average conceals the fact that value is violently concentrated. In the client work we've done, whether the business rents rooms, manages properties or sells contact lenses, a small slice of customers reliably carries a disproportionate share of future profit. When you build LTV the way an actuary builds embedded value, projected forward by segment, with explicit assumptions about frequency, margin and survival, that concentration stops being a vague intuition and becomes a map. You can see precisely which customers your future profit depends on, and what each assumption is worth.

Then you do what actuaries do next: you stress the assumptions. What moves the present value more, a 10% improvement in new customer volume or a 10% improvement in second-purchase rate among your top segment? I have run this sensitivity test many times. Retention assumptions win far more often than acquisition assumptions, and it usually isn't close.

Retention compounds. Acquisition resets to zero.

There's a second mechanism, and it's the one I find businesses most surprised by.

Customer survival is not a coin flip repeated each month. A customer who has stayed twelve months is meaningfully more likely to reach month twenty-four than a new customer is to reach month twelve. Loyalty is duration-dependent. The longer someone stays, the longer they tend to keep staying.

Play that forward and something powerful happens: your best customers are, by definition, the ones who survive longest, which means every improvement you make to retention acts most strongly on your most valuable people. The value doesn't just add up. It concentrates and compounds.

Acquisition has no memory. Every new campaign starts from zero, buys another blind draw, and pays full price for it. There is no mechanism inside acquisition that makes next month's acquisition cheaper or better because of last month's. Retention is the only side of the business where this month's work makes next month's economics structurally better.

This is the difference between effort and systems, and it's why I built CMPND around compounding in the first place.

The market is closing the acquisition arbitrage anyway

Even if you reject everything above, the maths of acquisition is deteriorating on its own. Industry estimates put customer acquisition costs up around 60% over five years, and aggregated benchmark analyses show a further 40% to 60% jump between 2023 and 2025 alone, driven by ad auction inflation, privacy changes that broke targeting, and plain saturation.

Read that again. In much of consumer commerce, acquisition is no longer where profit happens at all. Research tracking e-commerce acquisition economics found that the average brand now loses around $29 on a new customer's first purchase, up from a $9 loss a decade earlier, and only turns profit when that customer comes back. The first purchase buys a ticket. The repeat purchase is the business.

What I actually do about it

The method I run with clients is simple to describe and demanding to execute.

Build the model first. Not a generic LTV, but a segmented present value of future profits: who your customers are, what each group is genuinely worth going forward, and under what assumptions. Then work out the true, fully loaded cost of acquiring into each segment, including the costs most founders forget to count.

Then stress it. Move each assumption and watch what the valuation does. The assumption that moves the number most is your biggest lever, and in my experience it is almost never the one on the roadmap.

Then act by segment. Your above-average customers get strategies designed to keep them: recognition, tiers, economics that reward the behaviour you want more of. Your below-average customers get a different playbook, moved toward profitability or acquired more cheaply next time by fixing what your acquisition targeting got wrong. And your acquisition budget gets a ceiling that's finally honest, because now it's set against what each segment is actually worth, not against a fictional average.

We've built versions of this for accommodation operators, property managers, retailers and an education business, and the pattern holds every time. The leverage was sitting in the existing customer base, priced by nobody, while the budget chased strangers.

The bucket

Let me be fair to acquisition before I finish it off. You cannot retain a customer you never won. Acquisition is the entry fee, and in the early life of a business, simple acquisition-led growth is exactly right. When you have forty customers, you don't need an embedded value model. You need customer number forty-one.

But businesses graduate, the same way people do. What served you at forty customers quietly starts costing you at four thousand, because by then the biggest pool of unpriced value in the company is no longer out in the market. It's sitting in your own base, and the simple playbook has no tools for it. The point of this essay isn't that acquisition doesn't matter. It's that the moment your base is big enough to have segments, the repeat rate overtakes the acquisition rate in what it's worth to you, and most businesses sail years past that moment without noticing.

Which brings me back to the bucket. Your acquisition rate tells you how fast water is entering it. Your repeat rate tells you whether you own a bucket at all.

One of those is a marketing metric. The other is the valuation of your business, expressed as a behaviour. A company that repeats holds every rand it earns and stacks the next one on top. A company that only acquires rents its revenue month to month, at prices that rise every year, from platforms that don't care whether it survives.

So yes, I'll hold the hard line. Your repeat customer rate is more valuable than your acquisition rate, because it's the number that carries information, compounds over time, and prices your future. If you only measure one thing next quarter, measure the percentage of last year's customers who are still paying you. Then ask what that number would be worth if it moved five points, properly modelled, segment by segment.

That question is usually worth more than the entire acquisition budget. It's the question I start every engagement with, and it's a good one to sit with even if you never speak to me.