The Challenge
Our client is a fashion retailer with serious omnichannel scale – physical stores backing up a busy online storefront, and a CRM database to match. Years of list growth had left them with something a lot of retailers will recognise: a large, valuable database that was quietly getting tired, with a big chunk of subscribers who’d stopped meaningfully engaging with email at all.
The instinct, understandably, was to send more. At their peak the team was pushing up to seven batch campaigns a week, betting that more touchpoints to a wider audience would offset the drop in clicks.
It didn’t. It made things worse.
More sends to a database full of disengaged contacts just meant lower click rates at higher volume. Inbox providers noticed. Deliverability softened, more email landed in spam or promotions, engagement dropped further, and the team sent even more to compensate. Classic downward spiral, and one we see constantly in retail email programs that have scaled list size faster than they’ve scaled relevance.
The brief: break the spiral, without giving up total clicks and the revenue email was driving.
What We Did
We stopped treating frequency and audience size as the levers to pull, and went after the actual problem: the brand wasn’t sending the right products to the right people.
So we built recency and affinity modelling for every customer in the database: a live, constantly evolving picture of what each person actually cares about right now, not just in general.
The model works by scoring a range of behavioural signals that indicate genuine product interest – things like product views, wishlist adds, and on-site searches – and weighting each one by how strongly it tends to predict real intent. A wishlist add says more than a passing browse. We then layer in recency, so a signal from this week carries more weight than one from six months ago, and interest in any one category is scored relative to everything else that customer is engaging with, not in isolation.
The result is a dynamic affinity score per customer, per category — one that updates itself as behaviour changes, rather than a static profile that goes stale the moment it’s built. That score became the engine behind every send, letting the team cut weekly volume while making each email far more relevant to the person receiving it.
The Results
Within six months, the pattern was clear: send less, more precisely, and engagement comes back.
- Click rate up 123%, as campaigns were built around what customers actually wanted to see, not what the calendar demanded.
- Weekly batch touchpoints cut from 6 to 3 — half the noise, none of the lost share of voice with engaged customers.
- Total clicks doubled year on year, despite — or because of — a much lower send volume. Fewer, sharper emails beat more, broader ones.
- Unsubscribe rate down 50%, as customers stopped getting irrelevant content at a frequency that had been pushing them out the door.
The detail that got the most attention internally: send volume dropped significantly while total clicks went up. That’s the spiral running backwards: less volume, more relevance, better engagement, healthier deliverability, and every remaining send working harder than the last.
Sending more and getting less?
If your batch sends are climbing and your click rates are falling, the fix usually isn’t more email but smarter targeting. Get in touch with The House of Email to find out what recency and affinity modelling could do for your program.
