The challenge
Wren & Marlow sold candles and home fragrance direct to consumers with a 68,000-person email list. Email produced 9% of revenue, which they knew was low, and every campaign carried a discount code because campaigns without one performed badly. That was treated as evidence that their customers were price-sensitive. It was actually evidence that eighteen months of discount-led sending had taught the list to wait. Open rates were healthy, but purchases clustered entirely around discount sends, and full-price weeks were dead.
What we did
Showed them the cohort data
We segmented customers by whether their first purchase used a discount. Discount-acquired buyers had a 41% lower repeat rate and a clearly lower average order value. The list wasn't price-sensitive by nature. It had been trained.
Built the flows before touching campaigns
Twelve automated flows including a replenishment sequence timed from actual median reorder intervals per fragrance category, not a generic 30 days. That flow alone now produces a seven-figure annual run rate at full price.
Replaced discounts with access
New releases, restocks of sold-out scents, and early access for repeat buyers. VIP segments got first look, not deeper discounts, which cost nothing and performed better.
Sunset a third of the list
23,000 subscribers hadn't opened in twelve months. Removing them felt like destroying an asset and immediately improved inbox placement for everyone else, which is where a meaningful part of the revenue gain came from.
The result
Email revenue grew 212% and email's share of total revenue went from 9% to 34%, on a list that was a third smaller. Repeat purchase rate rose 47% and contribution margin improved 29%, because the revenue was no longer arriving with a discount attached.
What result did Wren & Marlow get? +212% in email revenue over 12 months, completed in 2026.
“They told us to stop discounting and delete a third of our list in the same meeting. I said no to both. We ran it as a test on one segment because they wouldn't drop it, and the test made the argument for them.”
Figures reflect real engagement outcomes at the time they were measured. This is a demonstration build — the client name and identifying details above are illustrative.