The Retention Maths: Why 5 Percent Fewer Lost Customers Can Mean 25 to 85 Percent More Profit
Most owners I meet know their cost per lead to the cent and could not tell me, even roughly, how many customers they lost last year. That is the wrong way round, and it is quietly costing them a fortune. The number that moves profit the most is not the one at the top of the funnel where all the budget goes, it is the one at the bottom, the leak, the customers who bought once and never came back. Here is the uncomfortable maths I want to walk you through. A retained customer pays you again with almost none of the cost you spent to win them the first time, so a small drop in how many you lose does not nudge profit up a little, it lifts it a lot. The classic finding is that cutting defections by 5 percent can raise profit anywhere from 25 to 85 percent. That range is real and it is worth understanding, and by the end of this you will be able to work out your own version of it on the back of an envelope, and then a proper worked version once you have your real numbers. The lowest cost growth in your business is almost certainly hiding at the bottom of the funnel, not the top.
The number nobody in the room can answer
Whenever I sit down with an owner I like to ask two questions early. The first is what does it cost you to win a new customer. Almost everyone has an answer, sometimes to the cent, because that number sits on a dashboard and somebody watches it every morning. The second is how many customers did you lose last year. Silence. Blank look. A guess, usually way too low, followed by a slightly nervous laugh.
That asymmetry is the whole problem in one moment. We measure, obsess over and pour money into the top of the funnel, the winning of new people, and we barely glance at the bottom of it, the losing of the people we already won. And it turns out the bottom of the funnel is where the lowest cost growth in your entire business is quietly sitting, waiting for somebody to pay attention to it.
I want to show you the maths, because once you have seen how a small cut in defections compounds into a large lift in profit, you will never look at your marketing budget the same way again. This is not a motivational point about loving your customers, though you should. It is an arithmetic point about where your profit actually comes from, and it holds whether you run a shop, a studio or a software product.
The finding that started all of this
Back in 1990 two researchers, Frederick Reichheld and Earl Sasser, published a piece in Harvard Business Review with a title that says the whole thing, Zero Defections. They went through the books of company after company across very different industries, and they found a pattern that keeps showing up. Reducing your customer defection rate by 5 percent could increase profit by somewhere between 25 and 85 percent, depending on the industry.
Now, I want to be honest with you about that range straight away, because it gets quoted like scripture and it should not be. It is a real finding, and the direction of it is rock solid, but 25 to 85 percent is an enormous spread, and where your business lands inside it, or whether you land outside it, depends entirely on your margins, your repeat behaviour and your cost of winning a customer. Do not take the headline number to the bank. Take the shape of it, then model your own. By the end of this article you will be able to.
The reason the effect is so large, and so reliable in direction even when the exact figure wobbles, comes down to one simple truth about a retained customer that most businesses never sit with properly. Let me walk you into it slowly, because it is the hinge the whole argument turns on.
Why a kept customer is almost pure profit
Think about what it costs you to earn a euro of revenue from a brand new customer versus a euro from one you already have.
The new customer arrives with a bill attached. You paid for the ad, the click, the content, the sales time, the discount you dangled to get them over the line. By the time they have handed you their first order you may have spent most of the margin on that order just getting them through the door. First orders are often close to break even, sometimes worse. That is normal and it is fine, because the plan was never to profit on order one.
The kept customer arrives with almost no bill at all. You already own the relationship. They know you, they trust you, they have your details saved and your emails in their inbox. Winning their second, third and fourth order costs you a rounding error next to what the first one cost. So the margin on those later orders is not eaten by acquisition. It mostly drops through to profit.
Here is the first thing worth sitting with. Two businesses can have identical revenue and wildly different profit, purely because of the mix between revenue that arrived with an acquisition bill attached and revenue that arrived almost for free from customers they kept. Top line looks the same. Bottom line does not. Retention is not a customer service nicety, it is a margin structure. When somebody tells me their revenue is flat but their profit fell, the mix between won revenue and kept revenue is the first place I look, and it is usually the answer.
Putting a number on a customer
To make this concrete we need a way to value a customer, and the cleanest simple model I keep coming back to is the customer lifetime value formula from the marketing analytics literature, in particular the framework in Cutting Edge Marketing Analytics by Venkatesan, Farris and Wilcox. It says a customer is worth their margin in a period, scaled up by how long they tend to stay and discounted for the fact that future money is worth less than money today.
Written out it is this.
Where M is the gross margin a customer produces in one period, usually a year. r is the retention rate, the probability they are still a customer next period. And d is your discount rate, the annual cost of capital that turns future euros into today's euros.
The fraction is where the magic and the danger both live. When retention r is high, the denominator 1 + d - r shrinks toward zero, and the whole multiplier balloons, because a loyal customer is worth a big stack of future years, not just this one. When retention is low, the multiplier collapses, because there simply is not much future to count.
Let me run one customer through it so the machinery is clear. Say a customer throws off 400 euros of gross margin a year, your discount rate is 10 percent, so 0.10, and 70 percent of these customers are still around next year, so r is 0.70.
So at 70 percent retention this customer is worth about 700 euros over their life with you. Hold that 700 in your head, because in a moment I am going to change one single thing and watch what happens.
A quick word on the two inputs people get wrong
Before the tables, two warnings, because I have watched owners feed rubbish into this formula and get a lovely looking number out.
The first is margin, not revenue. M is gross margin, what is left after the cost of the goods and the direct cost of serving that order, not the headline price. If you plug in revenue you will overvalue every customer by whatever your cost of sales is, and you will happily overspend to acquire them. Use the money that actually reaches you.
The second is that retention is a rate over a defined period, not a vibe. It is the fraction of the customers you had at the start of a window who were still buying at the end of it. If you eyeball it, you will guess high, because the customers who quietly drifted away do not send a leaving note. They just stop. Measuring it properly almost always produces a lower and more honest number than the one in the owner's head.
The table that makes the point
I am going to keep the margin fixed at 400 euros and the discount rate fixed at 10 percent, and I am going to move only the retention rate, from 0.70 up to 0.90. Nothing else changes. Same customer, same margin, same cost of capital. The only thing that moves is how many of them come back.
| Retention rate | Lifetime multiplier | Customer lifetime value |
|---|---|---|
| 0.70 | 1.75 | 700 euros |
| 0.75 | 2.14 | 857 euros |
| 0.80 | 2.67 | 1067 euros |
| 0.85 | 3.40 | 1360 euros |
| 0.90 | 4.50 | 1800 euros |
Look at what that column on the right does. A customer worth 700 euros at 70 percent retention is worth 1800 euros at 90 percent retention. Same margin. Same business. You moved loyalty by 20 points and you more than doubled what the customer is worth. And notice it is not a straight line. The jump from 0.70 to 0.75 adds 157 euros. The jump from 0.85 to 0.90 adds 440 euros. The higher your retention already is, the more each extra point is worth, because the fraction is accelerating as r climbs toward one.
That is the second thing worth sitting with, and it is the real engine under the Reichheld and Sasser number. Retention does not add to customer value, it multiplies it, and the multiplier itself gets steeper the better you already are. A small cut in defections, moving r up by a few points, does not move profit a little. It compounds.
The same customer in a thin margin business
I can already hear someone thinking that 400 euros of margin is a fantasy for their world, so let me run the identical exercise for a business where each customer only throws off 150 euros of gross margin a year. Discount rate still 10 percent. Watch the shape, because the shape is what matters, not my numbers.
| Retention rate | Lifetime multiplier | Customer lifetime value |
|---|---|---|
| 0.70 | 1.75 | 263 euros |
| 0.75 | 2.14 | 321 euros |
| 0.80 | 2.67 | 400 euros |
| 0.85 | 3.40 | 510 euros |
| 0.90 | 4.50 | 675 euros |
The euros are smaller but the geometry is identical. Moving from 0.70 to 0.90 still more than doubles the value of the customer, from 263 to 675 euros. The multiplier column is the same column, because the multiplier does not care about your margin, it only cares about how loyal your customers are and what your money costs. This is the bit I love. The retention lever works the same whether you sell garden furniture or accountancy retainers. Only the size of the prize changes, never the direction.
What the discount rate is quietly doing
People skip over d because it feels like an accountant's fiddle, but it deserves a moment, because it sets how much the future is allowed to count. The discount rate is your cost of capital, the return you could get on that money elsewhere, or the risk that the future is less certain than the present. A higher d says future euros are worth less today, which shrinks the value of a long lived customer.
Here is the same 400 euro customer at 80 percent retention, with only the discount rate moving.
| Discount rate | Lifetime multiplier | Customer lifetime value |
|---|---|---|
| 0.05 | 3.20 | 1280 euros |
| 0.10 | 2.67 | 1067 euros |
| 0.15 | 2.29 | 914 euros |
| 0.20 | 2.00 | 800 euros |
It matters, but notice it moves the number far less violently than retention does. Doubling the discount rate from 0.10 to 0.20 knocks the customer down from 1067 to 800 euros, a real dent but not a collapse. Moving retention from 0.70 to 0.90 more than doubled the same customer. So if you only have the appetite to obsess over one input, obsess over retention. Pick a sensible, slightly conservative discount rate, write it down, and then leave it alone while you go to work on the lever that actually swings.
Retention, churn and how long a customer really lasts
There is a mirror image of retention that makes the stakes even more visceral, and it is worth having both in your head. If retention is the fraction who stay, churn is the fraction who leave.
And if a customer has the same chance of staying each period, the average number of periods they last is one divided by the churn rate.
That little formula turns an abstract percentage into a lifespan you can feel.
| Retention rate | Churn rate | Average customer lifetime |
|---|---|---|
| 0.70 | 0.30 | 3.3 years |
| 0.75 | 0.25 | 4.0 years |
| 0.80 | 0.20 | 5.0 years |
| 0.85 | 0.15 | 6.7 years |
| 0.90 | 0.10 | 10.0 years |
Read that last column slowly. At 70 percent retention your average customer is with you for a bit over three years. At 90 percent they are with you for ten. You did not change the product, the price or the market. You changed how many walk out of the back door each year, and their expected lifetime tripled. That is the same compounding as before, wearing different clothes. When you cut churn you are not saving a few orders, you are extending a lifespan, and every extra year of that lifespan carries margin that cost you nothing to acquire.
Two loops, one leaky, one tight
Here is the same idea as a picture, because I think it is the clearest way to see the trap most businesses are in. There are two ways to grow. You can spend to win customers who then leak away, so you spend again to replace them, forever. Or you can keep the customers you win, so each new one adds to a base instead of just plugging a hole in it.
On the treadmill, on the left, your spending never stops and your customer base never really grows, because every new customer you win is roughly cancelled by one you lose. You run hard and stay in the same place. On the loop, on the right, the customers you keep compound. Each one you retain is a customer you do not have to win again, which frees the budget you would have spent replacing them to go and win someone genuinely new, on top of the base rather than instead of it.
The cruel part is that the treadmill and the loop can look identical on a monthly revenue chart for quite a while. The difference only shows up in the profit, and in what happens the day you turn the ad spend down.
Watch the base grow, or fail to
Let me put numbers on those two loops so the gap stops being a metaphor. Imagine both businesses win exactly 1,000 new customers every year. The only difference is retention. One keeps 70 percent, the other keeps 90 percent. Here is how the total customer base builds over five years, starting from nothing.
| Year | Treadmill base at 70 percent | Loop base at 90 percent |
|---|---|---|
| 1 | 1,000 | 1,000 |
| 2 | 1,700 | 1,900 |
| 3 | 2,190 | 2,710 |
| 4 | 2,533 | 3,439 |
| 5 | 2,773 | 4,095 |
Same acquisition, every single year. By year five the loop is carrying nearly half as many customers again as the treadmill, and the gap is still widening. Keep the maths running and the treadmill settles at around 3,300 customers and never climbs past it, because losing 30 percent of a bigger base eventually eats the whole 1,000 you add. The loop keeps climbing toward roughly 10,000, because losing only 10 percent leaves far more room for the new arrivals to stack on top. Same spend, wildly different business, and the only dial you touched was the one at the bottom of the funnel.
Where the money is versus where the budget goes
So here is the A ha I really want you to leave with. Most small and medium businesses pour almost all of their marketing effort and money into the top of the funnel, the acquisition, because that is the exciting part, the part with dashboards and agencies and campaigns. Meanwhile the bottom of the funnel, the leak, gets a newsletter nobody has redesigned since 2019 and a vague hope that people will come back on their own.
But the maths we just did says the lowest cost growth you can buy is almost never another point of conversion at the top. It is a point or two of retention at the bottom. Winning a brand new customer means paying the full acquisition bill again. Keeping one you already have means sending a well timed reminder, making the second order effortless, giving them a reason to come back. The second thing costs a fraction of the first and, as the tables showed, moves customer value more.
This is why I get slightly evangelical about it with owners. If your next marginal hour and your next marginal euro can go either into widening the top of the funnel or into plugging the bottom, the maths says plug the bottom almost every time, at least until your retention is genuinely good. You are not choosing between growth and loyalty. Loyalty is the more efficient growth.
A worked mini case, the leaky shop
Let me pull all of it together into one small worked example, the kind I sketch on a whiteboard in a first meeting. Numbers are illustrative, so swap in your own, but the machine is real.
Picture a shop that wins 1,000 new customers a year. Each customer throws off 200 euros of gross margin a year. It costs 150 euros to acquire one, and the discount rate is 10 percent. Today the shop keeps 70 percent of its customers from one year to the next.
At 70 percent retention each customer is worth 200 times 1.75, which is 350 euros of lifetime value. Take off the 150 euros it cost to win them and each customer nets 200 euros of profit over their life. Across the year's 1,000 new customers that is 200,000 euros.
Now suppose we do the unglamorous work and lift retention by five points, to 75 percent. Nothing else changes. Same acquisition, same margin, same cost of winning.
| Scenario | Retention | Lifetime value | Acquisition cost | Profit per customer | Profit on 1,000 customers |
|---|---|---|---|---|---|
| Today | 0.70 | 350 euros | 150 euros | 200 euros | 200,000 euros |
| Five points better | 0.75 | 428 euros | 150 euros | 278 euros | 278,000 euros |
| Ten points better | 0.80 | 533 euros | 150 euros | 383 euros | 383,000 euros |
Five points of retention turned 200,000 euros of profit into 278,000, a lift of about 39 percent, from work that cost a fraction of a new acquisition campaign. Ten points nearly doubled it. And here is the subtle bit that makes it better than it already looks. Because the acquisition cost is fixed and sits between the customer and their profit, every extra euro of lifetime value from retention lands almost entirely on the profit line. The cost of winning them was already paid.
One honest caveat so you do not quote me the way people quote Reichheld. I moved retention in whole percentage points here to keep it readable, which is not exactly the same as the 5 percent reduction in defections in the original study. The direction and the compounding are the point. Put your own margin, your own acquisition cost and your own retention into the formula before you promise anyone a number.
The common mistakes I watch people make
Once owners get excited about this, they tend to trip over the same handful of things. Here are the ones I flag most often.
The first is measuring retention on revenue instead of customers, or worse, not defining the window. Retention only means something over a stated period and against a stated starting group. Pick your window, name the cohort, and be consistent, or you are comparing numbers that were never the same shape.
The second is celebrating an average that hides a bleed. A healthy looking overall retention rate can sit on top of a first order retention that is dreadful, where most of the people who try you once never come back and a loyal core carries the average. The first repeat is usually where the biggest leak is, and the average will happily hide it from you.
The third is spending on winback before fixing the reason people leave. Chasing lapsed customers with discounts while the thing that annoyed them is still there just buys you the same churn again at a worse margin. Fix the leak, then pour.
The fourth is treating retention as the marketing team's job. Most of what keeps a customer is the product, the delivery, the support and the second order being effortless, none of which sit in the marketing budget. Retention is an operations and product question wearing a marketing hat.
The fifth is ignoring the discount rate entirely and then believing a fantasy lifetime value that stretches ten years into a future you cannot see. Be a little conservative. A slightly modest number you trust beats a heroic one you secretly do not.
How to apply this in your business this quarter
None of this needs a data science department. It needs three things wired together, and they are all within reach of a normal business.
First, measure the leak, because right now most owners cannot answer my second question. Work out your retention rate honestly. Of the customers who bought from you in a given period, what fraction bought again in the next one. Drop that single number into the formula above and you have what a customer is really worth, and tracking it over time tells you whether the leak is getting better or worse. What you do not measure, you cannot improve, and almost nobody measures this.
Second, own the relationship you are trying to keep. You cannot bring a customer back if you do not have a durable record of who they are and what they bought, one that survives longer than a browser cookie. That means a system you control, a real customer record, not a rented platform that forgets the customer the moment they close the tab. It is the same reason I keep pushing owners toward a shop and stack they actually own, and it is exactly the kind of foundation my ecommerce platform work is built to give you.
Third, act on the number. Once you can see retention, small things move it. A reason to come back before they drift, a second order made effortless, a subscription for the thing they buy on repeat, a reminder timed to when they actually run out. Each of these lifts r, and every point of r compounds through the formula into real profit. This is the work I love doing with clients, turning a leaky funnel into a loop, and if you want a hand finding and pricing your own leak it is squarely what my marketing and analytics work is built around.
Do these three in order. Measuring before you can act feels slow, but acting before you can measure is how businesses spend a year on retention tactics and never learn whether any of them worked.
The one number to leave with
If you take a single thing from this, make it this. You almost certainly know what it costs you to win a customer, and you almost certainly do not know how many you lose. Fix that, because the second number is where your profit is hiding. A customer you keep pays you again for almost nothing, a small cut in defections compounds into a large lift in profit, and the classic finding of 25 to 85 percent more profit from 5 percent fewer defections, while you should model your own version rather than quote mine, points in a direction that is simply true. Stop obsessing over the cost of winning. Start measuring the cost of losing. The lowest cost growth in your business is at the bottom of the funnel, and hardly anyone is down there looking.