Back to Work

Case 01 Loyalty eCommerce · Fashion

Building a loyalty programme that drives retention, not discounts

An established tiered loyalty programme looked healthy on paper, but it was rewarding markdown-heavy, high-returning customers and doing little to change behaviour. This is how I rebuilt it around repeat purchase, engagement and profitability rather than points and discounts.

01

The Situation

The business already had an established loyalty programme with points, tiers and benefits, but in practice it behaved more like a passive discount scheme than a retention programme.

Customers were automatically enrolled, points were largely earned through purchases, and there was very little communication encouraging members to engage with the programme after joining. The programme existed, but it wasn't doing enough to actively change customer behaviour.

I wanted to understand whether loyalty was genuinely encouraging our best customers to shop more often — or simply making purchases cheaper for customers who would have bought anyway.

02

What I Found

The biggest issue wasn't the loyalty mechanics themselves. It was who we were rewarding and what behaviour we were encouraging.

59%

of our highest-tier customers' purchases were already made on markdown — while those customers received the most expensive benefits.

~40%

of all product returns came from customers within the two highest tiers.

This created an uncomfortable commercial dynamic: some of the customers receiving the greatest loyalty investment were also among the most expensive customers to serve.

Tier-level analysis

Markdown share

Products sold discounted

Level 449.2%
Level 328.2%
Level 222.8%
Level 117.9%

Return rate

Returns as share of tier sales

Level 433.1%
Level 330.6%
Level 224.0%
Level 128.7%

Benefit cost

Avg. annual per member

Level 4£44.90
Level 3£25.10
Level 2£11.90
Level 1£4.30
Level 4 is the top tier, highlighted across all three measures: it carried the heaviest markdown dependency, the highest return rate and the largest benefit cost per member. Each panel is scaled independently — values are labelled in full.

There were other structural problems. Customers were automatically enrolled into loyalty, which meant a large proportion of the membership base had made only one lifetime purchase. Movement between tiers was slow, benefits were relatively similar and heavily purchase-led, and there were few reasons to interact with the programme between transactions.

CRM wasn't helping enough either. There were only a small number of automated loyalty journeys and almost no regular broadcast activity reminding customers about their points, status or benefits.

The programme was rewarding transactions, but wasn't doing enough to create loyalty.

03

My Thinking

The easy response would have been to make the programme more generous. More points. Bigger discounts. Better tier rewards. I felt that would make the underlying problem worse.

The opportunity was to move loyalty away from "spend money, receive discount" and towards a programme that encouraged a broader set of valuable behaviours: making a second purchase, progressing through tiers, engaging with the brand and ultimately building a more profitable customer relationship.

Not every high-spending customer should automatically be considered a high-value customer.

Returns, markdown dependency and the cost of benefits needed to sit alongside revenue when determining customer value. I therefore structured the redesign around three questions:

  • How do we make loyalty something customers actively want to participate in?
  • How do we use it to create more reasons to return without relying on discounts?
  • How do we make sure the economics still work for the business?

The objective wasn't to remove financial rewards completely. Discounts still had a role. The difference was making them one part of the proposition rather than the proposition itself.

04

What Changed

1 Qualification became intentional

I moved away from automatically treating every account holder as an active loyalty member. Customers would enter the rewards programme after making their first purchase, creating a clearer distinction between someone who had simply created an account and someone who had started a relationship with the brand.

From there, CRM could focus immediately on the next important behaviour: getting that customer to purchase again.

2 We expanded what loyalty could reward

Points had previously been driven almost entirely by spend. I developed a broader engagement model where customers could also be recognised for behaviours such as downloading the app, enabling notifications, completing their profile, submitting reviews or UGC and engaging with selected brand activities.

Limits were built into repeatable actions to prevent points farming. The principle was simple: reward behaviours that deepen the relationship, not activity for activity's sake.

Ways to earn points

Place an order £1 = 1 point
4 Every 4th order 200 points
Download the app 300 points
Newsletter sign-up 200 points
Follow us on socials 200 points
Product review 100 points
Review with photo 150 points
Birthday reward Wait and see!
The expanded earning model — points attached to behaviours beyond spend, with limits on the repeatable actions.

3 Loyalty became part of the CRM lifecycle

Rather than leaving customers to discover benefits themselves, I mapped a new set of automated journeys around the moments where loyalty could influence behaviour. These included post-purchase education, points reminders, approaching-the-next-tier messages, tier upgrades, expiry warnings, downgrade grace periods, birthday rewards and targeted lapse-prevention journeys for high-value members.

I also introduced loyalty into the BAU CRM calendar through points reminders, member-only content, early access, priority back-in-stock communications, selected double-points events and exclusive experiences.

The intention was to turn loyalty from something sitting quietly in an account page into an ongoing retention conversation.

Redesigned loyalty lifecycle

01

Onboarding

First purchase and first tier

Welcome Journey

2 days after first purchase

New Tier Welcome

2 days after a new tier is reached

02

Retention

Keeping members active between purchases

Post-Purchase Journey AI

Probability to convert in the next 30 days

Points Reminder

45 days of no activity

Next-Tier Progression

Within 150 points of the next tier

Loyalty-Branded Retargeting

Abandoned cart, browse and wishlist

Birthday Journey

14 days before the birthday date

Lapsed Prevention AI

Likelihood to lapse in the next 30 days

03

Tier movement

Expiry, grace and downgrade

Tier Expiry Warning

30 and 7 day countdown to status drop

Tier Downgrade Grace

30 days before downgrade

Tier Downgrade

1 day after downgrade

04

Reactivation

Winning back lapsed members

Reactivation Journey AI

Missed average days between purchases

Channels across every journey Email · SMS · WhatsApp · Web layer · Mobile push
Twelve automated journeys mapped across the four lifecycle stages. Each is triggered by a loyalty behaviour rather than a calendar date; those marked AI are driven by propensity models rather than fixed rules.

4 We introduced profitability safeguards

One of the more important changes was recognising that spend alone couldn't determine loyalty value. The redesigned model included the ability to reduce or restrict benefits where customer behaviour — particularly consistently high return rates — made the relationship commercially unsustainable.

I also reviewed the tier structure itself, including the spend required to reach the highest level, while making the difference between tiers more meaningful through access, recognition and experiences rather than simply increasing discounts.

05

The Impact

The work fundamentally changed how we thought about the role of loyalty.

Instead of asking "What benefits should each tier receive?", the conversation became "What customer behaviour are we trying to create, and what is that behaviour worth to the business?" That led to a proposition built around repeat purchase, engagement and customer value rather than points accumulation alone.

It also gave CRM a much larger role in loyalty. The redesigned lifecycle introduced touchpoints across onboarding, progression and reactivation, while the monthly CRM calendar gave members regular reasons to understand and use their benefits.

Importantly, profitability became part of the programme design from the beginning rather than something reviewed after rewards had already been issued.

At the point of redesign, the next measurement phase was therefore deliberately focused on:

Second-purchase rate Tier migration Repeat purchase Reward utilisation Return-adjusted customer value Incremental revenue from loyalty members

— rather than simply programme membership or points issued.

Impact of the redesigned proposition

Customer behaviour

12%

Second purchase rate

First order through to second

40%

Reward utilisation

Earned rewards actually redeemed

15%

Purchase frequency

Orders per member

Programme & commercial value

20%

Tier progression rate

Members moving up a tier

10%

Full-price purchase share

Shift away from markdown

18%

Loyalty benefit cost

Per member — lower is better

Movement against the measures the programme was rebuilt around. All six moved in the intended direction — benefit cost is a reduction, achieved while behaviour and customer value rose.

Want to discuss a project?

Get in touch ↗