Before — the standalone referral page

Fig. 02The standalone referral page, as inherited.

Ray-Ban partner campaign imagery: friends at a golden-hour dinner table, one wearing Ray-Ban sunglasses. AI-generated, art directed by Michael James Clarke — Fig. 03Ray-Ban partner campaign. Imagery is AI-generated — I art directed concept, styling, composition, and colour.

Ray-Ban partner campaign imagery. AI-generated, art directed by Michael James Clarke — Fig. 10Ray-Ban campaign — the accidental-snapshot direction: warm, imperfect, mid-conversation.

Ray-Ban partner campaign imagery: woman in Ray-Ban glasses at an outdoor table. AI-generated, art directed by Michael James Clarke — Fig. 11Ray-Ban campaign — product worn in the moment, never staged on a shelf.

After — referral surfaced in context

Fig. 04Referral surfaced in context after the rebuild.

Share moment — native share sheet in the app

Fig. 05The native share moment.

Friend landing page

Fig. 06The friend landing page.

Refer a Friend campaign imagery: two friends sharing a meal at sunset. AI-generated, art directed by Michael James Clarke — Fig. 12Campaign photography — a great week of cooking, shared. AI-generated, art directed end to end.

Refer a Friend campaign imagery. AI-generated, art directed by Michael James Clarke — Fig. 13Campaign photography — the moment worth telling a friend about.

Refer a Friend campaign imagery: a couple looking at a phone together at a truck tailgate. AI-generated, art directed by Michael James Clarke — Fig. 14Campaign photography — the share moment itself.

Results — referral volume against referral quality, by cohort

Fig. 08Referral volume against referral quality, by cohort.

Reward tracking — invite status and credit

Fig. 07Reward tracking: invite status and credit.

The referral system, end to end

Fig. 09The referral system, end to end.

Hero — the referral ecosystem across touchpoints

Fig. 01The referral ecosystem: email, app, and the box itself.

Refer a Friend · HelloFresh · 2026

One referral system, rebuilt once. Nine brands plug in.

HelloFresh treated referral as a page. I rebuilt it as infrastructure every brand could plug into.

One component system. One CRM template. One measurement schema. Built that way from week one, so the Q3 global rollout is a configuration job, not a rebuild.

Referral volume fell 24% year over year. That was the point — the people who arrived converted better, stayed longer, and cost less to acquire.

Role: Lead Product Designer
Team: Share Squad
Scope: Strategy → end-to-end design
Markets: US + Canada, rolling out group-wide
Reuse: One system, 9+ brands · configure, don’t rebuild
Status: Shipped · documented as designed

−24%

YoY conversion volume

+1.4pp

US conversion rate

+8.3%

Friend net CCV

−5%

Unprofitable conversions

02 — Reframing the Problem

The referral page wasn’t the only experience shaping referral behavior.

My strongest hypothesis was that referral performance was downstream of the broader customer experience. If customers genuinely wanted to recommend HelloFresh, the act of sharing could be extremely simple. That meant the UX opportunity was larger than optimizing one page.

“If the product experience is strong enough, the RAF page could be a white screen with an email field.”

03 — Designing for Moments, Not Channels

The core HelloFresh experience happens beyond the app.

The team was heavily focused on the app experience. I pushed us to look at the full service journey instead. HelloFresh’s primary customer experience is the meal that arrives at someone’s doorstep—not just the interface they use to manage it.

04 — Rebuilding the Experience

From a referral page to a referral ecosystem.

I designed the program end to end: sender experience, friend landing, in-app touchpoints, native share mechanics, CRM direction, visual language, and physical packaging concepts—treating referral as one connected UX system.

05 — Results

Volume fell. Quality improved.

The first clean read was not a simple growth story. Fewer invites were being sent, but the invites that remained converted better and produced more valuable friends.

06 — Impact & Reflection

The rebuild created a stronger UX foundation—and clarified the next problem to solve.

What changed

The program moved onto a more modern foundation, gained cleaner measurement, simplified CRM, introduced personal sharing mechanics, and created a playbook sister brands could eventually adopt instead of maintaining fragmented referral programs market by market.

The most valuable outcome was clarity: we could separate visibility, invite intent, conversion, and friend quality instead of treating RAF as one blended number.

What I still believe

A referral experience cannot be separated from the product experience that creates advocacy. The strongest referral strategy starts upstream—in the moments customers receive, use, and enjoy the product—then makes sharing those moments easy.

Design’s role is to connect those moments into one coherent experience: reducing friction at the share point while also creating relevant physical and digital touchpoints throughout the journey.

Great referral UX starts before the share button—and makes the moment of sharing feel effortless when customers get there.

Takeaway

Case study takeaway: RAF taught me that referral is not a page, it’s a system. The strongest growth loops are designed upstream of the share button, in the moments that make someone want to recommend a product in the first place.