Why your retention rate is probably lying to you
You pull up your dashboard. Retention is up 4% month over month. You feel good about it for about ninety seconds. Then you notice something: the same customers you "retained" are the ones who've been with you since the beginning, and almost nobody has joined them. Your retention rate went up because your acquisition went down. The percentage improved while your business got worse.
This is the trap I fell into for most of my first two years running a small subscription product. I was optimising a number, not a behaviour. And the number was flattering me.
Here's what I want to argue in this article, and I'll be upfront that it's an opinion: most advice about improving customer retention rates is useless because it treats retention as a marketing metric rather than an operational outcome. Loyalty programs don't retain people. Neither do discount codes. What retains people is that the thing they bought keeps being worth more than what they paid — and that when it breaks, someone fixes it fast.
Everything below is built on that premise. Some of it comes from mistakes I made. Some of it comes from a four-month experiment that cut my churn from 6.1% to 3.4% per month, which I'll detail honestly, including the part where I nearly gave up on it.
Key takeaways
- Your retention rate can improve while your business shrinks — always read it alongside new-customer volume.
- Retention is decided in the first 30 days far more than in the twelfth month. Most churn is a failure of onboarding, not loyalty.
- Segment before you act. A single "retention strategy" applied to everyone helps almost no one.
- Silent churn — customers who stop using you but never cancel — is often larger than visible churn.
- Fix your worst moment before adding any new feature. One broken step churns more people than ten missing ones.
- Measure cohort by cohort. Blended averages hide everything that matters.
How to improve customer retention rates: a working definition first
Customer retention, stripped of jargon, means this: of the people who bought from you in one period, how many are still buying in the next? That's it. Everything else — loyalty tiers, points, referral programs — is a tool you might use to move that number.
The retention rate formula, and why it misleads
The standard formula is straightforward. Take the number of customers at the end of a period, subtract the new ones you acquired during it, and divide by the number you started with. Multiply by 100.
Run it once and you'll see the problem immediately. If you lost 400 customers and gained 400, your retention rate looks fine. But you paid acquisition costs to stand still. And acquired customers are almost always less loyal in their first months than the ones who just left, so the underlying quality of your base dropped even though the percentage didn't.
I now track three numbers side by side, and I'd argue you should too:
- Retention rate — the headline figure.
- Net revenue retention — same customers, this period vs. last. This one catches downgrades that the headcount figure misses entirely.
- New customer volume — because a rising retention rate with falling acquisition is a warning, not a win.
Combined, those three tell you whether you're actually keeping people or just papering over the exits.
Which numbers actually predict churn
Retention rate is a lagging indicator. By the time it moves, the decision was made weeks earlier. What I watch instead are three earlier signals: how many days since a customer last logged in, whether they've completed the core action that made them buy in the first place, and how many support tickets they've opened. Someone with four tickets and no logins in three weeks is gone. The headline number just hasn't caught up yet.
The silent churn problem nobody measures
Here's the thing that changed how I think about all of this. When I finally sat down and looked at my own data properly, I found that roughly a third of my "retained" customers had stopped using the product entirely. They hadn't cancelled. They were just... there. Paying, in some cases, out of inertia.
That's silent churn, and I'd argue it's the most under-measured number in retention work. Visible churn is the person who hits cancel. Silent churn is the person who's already gone but hasn't told you yet.
How to find your silent churn
Pick a definition of "active" — one meaningful action within a set window, ideally tied to the reason people bought. Then count how many paying customers fall outside it. When I did this, I got a number 40% larger than my visible churn. That's not a rounding error. That's a business.
The fix isn't a re-engagement email campaign blasted at everyone. It's a real question: why did they stop? And the honest answer, in my case, was that they'd hit a specific wall in week three and nobody had noticed.
A step-by-step plan that actually moves the number
Most retention advice gives you a list of tactics. What's usually missing is the order — and order matters enormously, because doing these out of sequence wastes months.
Here's the sequence I'd use. It took me a while to arrive at it, and the first version I tried was wrong in a specific way I'll come back to.
- Audit your worst moment. Find the single step where the most customers disappear. Not the average — the cliff. This is always step one, because adding features on top of a broken foundation just means more people fall through.
- Segment by value and by age. A customer in their first month needs something completely different from one in year two. Treating them as one group is the mistake I made for eight months.
- Define a small set of leading indicators. Three or four, tied to real usage, not vanity metrics.
- Test one intervention at a time. This is the boring part. Do it anyway.
- Measure by cohort, not in aggregate. Blended averages hide everything.
Where I got the order wrong
I started with the loyalty program. Six weeks of work, a points system, a nice little dashboard for it. Usage among the target group rose about 9%. The churn rate didn't move at all.
Of course it didn't. The people I was trying to retain were stuck on a step that came before any loyalty program could matter. I'd built a reward for finishing a race most of them never reached the starting line of. That was an expensive lesson in sequencing.
Once I went back and fixed the actual blockage — a setup step that took eleven minutes and could be done in two — the same cohort's 30-day retention went from 61% to 78% over about ten weeks. No points. No rewards. Just removing friction.
Retention strategies ranked by what they really do
Not all retention strategies are equal. Some address the root cause of churn. Others just delay it. Here's how I'd sort them.
| Strategy | What it actually does | When it works | Typical effort |
|---|---|---|---|
| Fixing onboarding friction | Removes the reason people leave early | Almost always; highest leverage | Low to medium |
| Proactive support | Catches problems before they become exits | When you have clear usage signals | Medium |
| Personalisation | Increases relevance, so the product fits better | Once you have enough data per customer | Medium to high |
| Loyalty and rewards programs | Delays the decision to leave | Only after retention fundamentals are solid | Medium |
| Discounts and win-back offers | Buys a short extension, rarely a relationship | As a last resort for specific segments | Low |
Read that table top to bottom and you'll see my bias. The cheap, unglamorous fixes sit at the top. The ones that get all the marketing attention sit near the bottom, doing the least structural work.
What about retention in retail and e-commerce?
Same logic, different mechanics. In physical retail, the equivalent of onboarding friction is the first return. If a customer's first return is painful, they don't come back — and returns are where most retailers quietly lose people. In e-commerce, it's usually delivery reliability and whether the product matches the listing. I've watched a small online store cut repeat-purchase churn simply by changing the photos on three of its top products to be less flattering and more accurate. Fewer first-time buyers, far more second-time ones.
Choosing your budget, and finding out you can't do everything
Here's the question I get asked most, and the one most articles dodge: where should the money go — acquisition or retention?
My honest answer is that the framing is wrong. You don't divide a fixed budget between them, because they aren't competing for the same work. Acquisition spends on reach. Retention spends on fixing the product experience. If your onboarding is broken, every dollar of acquisition is buying a customer who'll leave in ninety days.
A more useful question is: what's the weakest link right now? If you're losing more than a quarter of new customers in the first month, don't touch acquisition until that's fixed. If your first-month retention is solid, then more acquisition makes sense. There's no universal percentage split, and anyone giving you one is guessing.
The part I still struggle with
Attribution. Even with clean cohort data, I can't reliably tell you which specific tactic caused which improvement. When retention went from 6.1% to 3.4% monthly churn over four months, I'd changed three things. I like to believe the onboarding fix did most of the work, but I genuinely don't know. Anyone who tells you they've isolated it perfectly is either running a much bigger operation than most of us, or not looking closely enough.
What I do trust is direction. The churn trend moved down and stayed down. The cohort curves flattened. Whatever combination caused it, the change survived contact with reality, which is more than most of my earlier experiments did.
The question that matters more than any metric
After enough months of this, I've come to think the retention rate isn't really a marketing number at all. It's an opinion your customers hold about whether you're worth the trouble this month. No dashboard shows that directly.
So the question I'd leave you with isn't "what's my retention rate." It's this: if you called your ten longest-standing customers today and asked why they stayed, would you hear about the product, or about the fact that you've always answered the phone? For a while, mine was the second one. That's a compliment to my team and a warning about my product.
Improving retention rates starts the moment you're willing to hear that answer without defending yourself.