top of page
2 days ago
3 min read

Why did our fashion brand’s repeat rate fall while returning sales grew?

By Trayan Hristov · Founder, Thrivelia · 19 November 2026

Returning sales and repeat rate can move in opposite directions. A larger base of existing customers can spend more this quarter while a smaller share of this quarter’s new buyers comes back. Before I ask the CRM team to fix anything, I’d establish which of those things happened.

Put the same customers on the same clock

Start with first-time buyers grouped by the month of their first paid order. Count a second paid purchase within 90 days, and give every customer the full 90 days to qualify. Remove replacements, split shipments and test orders. Write down how you handle a fully refunded first order. If you acquired and served that person, I’d keep them in the acquisition cohort and flag the return. Dropping them can make an expensive acquisition channel look healthier.

Split each cohort into groups with a plausible difference in behaviour. Full-price basics, promotional purchases and occasionwear are a better starting point for a clothing brand than ten finely sliced audience personas. A customer buying a pack of everyday tees has another reason to come back relatively soon. Someone buying a dress for a wedding may be perfectly happy months later.

Change the weights before you change the programme

Here’s a fictional store. In spring it acquired 800 basics buyers and 200 occasionwear buyers. Within 90 days, 30% of basics buyers and 10% of occasionwear buyers bought again. That’s 260 out of 1,000, or 26%. In summer the mix flipped to 400 basics buyers and 600 occasionwear buyers. The repeat rates inside the groups stayed exactly the same. The overall figure fell to 18%.

First purchase

Spring buyers

Summer buyers

90-day repeat in both

Basics

800

400

30%

Occasionwear

200

600

10%

All customers

1,000

1,000

26% spring, 18% summer

Apply spring’s 80/20 buying mix to summer’s repeat rates. You get 26% again. In this example the shift in product mix accounts for all eight percentage points. In your store, compare the actual result with that fixed-mix result. Put counts next to the percentages. A small collection can move several points on a handful of orders.

Give the right person the next decision

If basics hold steady and occasionwear becomes a larger part of acquisition, merchandising and paid media belong in the review. Check the margin and acquisition cost of the dress before deciding that its slower repeat cycle is a problem. If basics buyers themselves stop returning, inspect availability in their original size, returns and the elapsed time to a second kept order. That gives CRM a much more specific brief.

Our perfectwhitetee case tracks changes in the owned-channel programme over time. For this question I’d put those channel figures beside the customer cohorts. A flow can get more efficient while the assortment attracts a different kind of first-time buyer. Shopify’s customer reporting definitions.

The five columns I’d take to the weekly meeting

Show prior repeat rate, current repeat rate, prior buyer mix, current buyer mix and eligible buyers for every segment. Add a second line for the fixed-mix result. If a new category only exists this season, show it as a new line and calculate the like-for-like comparison on shared categories. Otherwise a product launch with no historical baseline quietly determines the answer.

To turn the diagnosis into an operating target, multiply the achievable change by the eligible population. Improving repeat among 400 basics buyers from 30% to 32% means eight extra second buyers, before allowing for uncertainty. Compare the likely margin on those eight buyers with the cost of the intervention.

Ask ChatGPT or Claude about your own numbers

Use my aggregated acquisition-month, first-product group, eligible buyers and 90-day repeaters. Show the raw rates and reweight the new cohort to the old mix. Which group actually changed, and where are there too few buyers to tell?

Related reading

 
 
 

Comments


bottom of page