Most Personalization Isn't Personal — So It Doesn't Bring Customers Back

A few years ago, at a large membership retailer where I led CRM and membership analytics, I sat through a review where the team was proud of a campaign. It was called personalised. I asked what made it personal.

The answer was a segment. Customers who had bought a certain product got an email about a related product. Everyone in the segment got the same email, on the same day, with the same offer. The only thing that changed was the first name at the top.

We had not personalised anything. We had sorted people, then stopped. Sorting, on its own, does not bring a customer back.

That is the question behind this post: what's the difference between segmentation and personalization? Most teams I meet use the two words as if they mean the same thing. They do not, and the gap between them is where a lot of repeat buying quietly leaks away.

What the two words actually mean

Segmentation means splitting your customers into groups who share something: people who bought in the last 90 days, people who bought twice, people in one city. It is useful. It stops you sending one message to everybody.

Personalization means changing the message for the individual person inside that group, based on what you know about them. Not just their name. What they bought, how much of it, how fast they use it, whether they open your emails, and when they are likely to need you again.

Here is the test I use now. Take the customer's name off the email. Would the recommendation still change from person to person? If everyone in the segment would still get the same product, the same offer, at the same time, that is segmentation. It might be good segmentation. But it is not personalization.

That matters for retention. A first purchase can come from a good ad or a good price. A second purchase comes from timing and fit: did you show up when they actually needed the next thing? A segment cannot answer that. Only the person's own pattern can.

How teams do this now, in practice

The better-run small teams I have worked with keep their segments simple. Then, inside each segment, they let a few individual facts change the message: what the person needs next, when they need it, and where they will actually see it. A product that lasts 30 days and one that lasts 90 days do not share a send date. Some customers open email, some only tap a text, and sometimes the personalised decision is to hold back, because the customer is not due yet.

None of that needs a huge system. Most email tools already let you branch on purchase history, on how long a product lasts, and on whether someone opened your last few messages. What it needs is someone deciding, before the campaign goes out, which individual facts are allowed to change the message. Without that decision, the tool just sends the segment email faster.

The mistake is stopping there and calling the work personal.

One worked example

Take a small skincare brand with a segment of 2,000 customers who each bought the same face cream once. The next campaign goes only to that segment, so the team calls it personalised. Everyone gets the same email at day 30: a reminder to reorder, with 15% off.

Look at two people inside it. Customer A bought the small jar, uses it every morning, and it lasts about 35 days. She opened the last two emails. At day 30, a plain reminder makes sense. She is nearly out, and she is paying attention.

Customer B bought the same cream as a gift. She has never opened an email from the brand, and the cream was never for her. At day 30, she does not need a refill.

Send both the same 15% off email and Customer A may buy — and you will never know if the discount did it or the reminder did. Customer B ignores it, or unsubscribes, because it was never relevant to her.

Now scale it. Say 600 of those 2,000 people are like Customer B in one way or another: they bought a gift, or the large jar that lasts 90 days, or they bought once and never engaged again. None of them are due for a refill at day 30. Sending all 600 a 15% discount on a $50 order gives away $7.50 each, or $4,500 in total, for sales many of them were never going to make that week anyway. The figures are illustrative, not a real brand's results, but the shape is what I have seen again and again: a discount paying for timing that was wrong.

The fix is not a new platform. Split that segment once, by what changes the next step: who bought the small size and is nearly out, who bought the large size and is not due for weeks, who bought as a gift. The first group gets the reminder at the right time, at full price. The second group gets nothing until they are due. The third gets a different message, if they get one at all.

That is personalization. The message changed because the person changed, not because the segment label changed.

What changed

A few years ago, this kind of per-person timing was hard for a small team. It meant manual lists and someone remembering to pull a new list every week. That part has changed. Most tools a small brand already pays for can now branch a message on purchase history, engagement, and timing, without a separate project. What has not changed is the habit. It is still faster to build one email per segment, add first names, and call it personalised. The tools got cheaper. The thinking did not get automatic.

There is a second change, about trust. A refill reminder that lands the week a customer runs out feels helpful. The same reminder, three weeks early, feels like noise, and noise makes people stop opening — which loses the cheapest way to bring them back.

What I would do on Monday

Pick one campaign you currently call personalised. Take the name off five emails from it, at random. If you cannot tell the recipients apart from what is inside — same product, same offer, same timing — you have a segmentation campaign, not a personalization campaign. That is fine. Just name it honestly, and then fix one thing.

Fix the timing first. Split that audience by when they are actually due for the next purchase: what size they bought, how long it lasts, whether it was a gift. Send to the people who are due, hold the rest, and watch who actually buys again over the next month, and at what price — not just opens.

Personalization is not a fancier segment. It is a decision you make for one customer, based on what you know about that customer, that changes what they receive. When that decision is about the next purchase — the right product, at the right time, on a channel they use — it becomes the reason they come back.

This is the kind of work I now do with B2C teams: not more segments, but messages where individual facts change the next step.

Sources: Cardinal Path, Personalization vs. Segmentation: Understanding the Difference & Activating the Benefits; MoEngage, Customer Segmentation and Personalization: How They Differ; Latent View, What Is Marketing Personalization? The worked example figures are illustrative, drawn from the author's practice, not reported results from a named brand.