If you want to make your e-commerce marketing more efficient, there’s one place you should start before anything else – and it’s not your creatives, your audience size, or even your platform strategy…
It’s your customer base.
Specifically, how recently each person last bought from you.
This isn’t a nice-to-have segmentation model. It’s the foundation for deciding how much to spend, what offers to run, and who’s even worth chasing in the first place.
Start with Recency
Break your customers down by how long it’s been since their last purchase. Not revenue. Not product category. Just time.
For example:
- Bought in the last 30 days
- Bought 6 to 12 months ago
- Bought 2 to 4 years ago
- Bought 5+ years ago
Once you’ve grouped them, tailor your marketing based on how long it’s been since they last showed up.
Think about it like this: someone who hasn’t bought in three years isn’t just on a break – they’ve probably forgotten you exist. To bring them back, a 10 percent off email won’t cut it. But someone who last purchased 30 days ago? They don’t need a discount – they need a reason to come back soon.
Find the Cut-off Point
Here’s where things get interesting. You need to push this segmentation until you hit the point where your return on ad spend goes negative.
That’s your hard stop.
For one brand, that drop-off point might come after 6 months. For another, it might be 3 years. But once you find it, you’ve drawn a line in the sand. You now know which customer segments are still worth investing in – and which ones you should stop chasing.
This kind of clarity saves you thousands.
Why This Works
Two big reasons:
- You protect your margins
By holding back discounts from recent buyers, you stop training full-price customers to wait for deals. The longer you can keep them in the habit of paying full price, the more profit you retain. - You avoid throwing good money after bad
Some customers are gone for good. No amount of retargeting or email nurturing will bring them back. By identifying when they’re truly “done,” you can cut them loose and reallocate that spend toward people who still have buying potential.
Add Frequency and You’re in Business
Want to go deeper? Add a second layer to your segmentation: purchase frequency.
A customer who bought once a year ago is not the same as someone who made five purchases in the same time frame.
Those two variables – recency and frequency – are your best predictors of future value. Not demographic data. Not social media behaviour. Just time and repetition.
Real-World Example
We did this with a skincare brand that had over 400,000 customers in their database. Once we mapped recency and frequency, we discovered that anyone who hadn’t bought in 18+ months and had only ordered once was a sunk cost. Campaigns aimed at that group consistently delivered negative ROAS.
But customers who hadn’t bought in 9 months but had purchased 3+ times? Those were gold. We rebuilt their winback strategy around that insight – different offer structure, better message match, and a separate budget. That campaign hit a 4.2x ROAS in the first 45 days.
All we changed was who we spoke to, and how.
