Arket — 2022
Arket's conversion rate was two percent.
That number was not the problem.
It was the symptom.
The actual problem was structural: the brand had no mechanism to exist in a user's life between visits. Every session started from zero. There was no continuity, no accumulation, no signal that the brand remembered anything about the person standing in front of it. For a brand that positioned itself as considered, unhurried, quality-first — a brand that asked customers to slow down and choose well — the digital experience was doing the opposite. It asked for a transaction before it had offered a relationship.
The attempted fix made it worse. A ten percent discount for account creation failed not because the offer was insufficient but because it was incoherent with the brand. Arket doesn't compete on price. It competes on the value of the object and the experience of choosing it. A coupon said: we need you to commit. The brand said: take your time. Those two things cannot coexist at the same entry point.
I was brought in to diagnose why people weren't returning and to build the strategic framework that would change the brand's relationship with its audience — at scale, across the H&M Group.
I designed a research program pairing behavioral analytics with ethnographic observation. I analyzed clickstream data from thousands of sessions — where people slowed down, what they zoomed in on, which moments preceded abandonment. Then I sat with users in their homes and watched them shop in real time.
The behavior was consistent. A woman in Stockholm found a linen dress at midnight, photographed it, and closed the tab. "I'm not ready to buy," she said. "But by tomorrow it'll be gone from my head." Another encountered a sold-out jacket and moved on without hesitation. "There's no way to know if it'll come back, so why bother?" A third named the brand experience exactly as it was: "It's like shopping in a store where nothing remembers you."
That sentence was the brief. Not a UX problem. A brand memory problem.
The insight I brought to the first cross-functional review reframed the entire conversation. We weren't failing at conversion. We were failing at continuity. Users arrived with genuine intent — strong, specific, considered intent — and found no place to put it that didn't require an immediate purchase decision. The brand was asking for trust before it had demonstrated presence. It was asking people to commit to a relationship the brand itself hadn't shown up for yet.
The strategic reframe was this: stop asking users to log in for a discount they don't value, and start giving them something worth protecting.
Brand loyalty is not built at the moment of purchase. It is built in the quiet moments before — when someone is still deciding whether a brand is worth remembering. If those moments have nowhere to go, the loyalty evaporates. The solution was to build infrastructure for those moments: a way for users to express and accumulate intent before they were ready to act on it.
I led the strategic definition of Favourites — a feature allowing users to save items without logging in, with no form, no commitment, and no friction. It persisted across sessions through browser storage, letting people build collections anonymously until they had something worth making permanent.
Early testing exposed a gap I had anticipated from the research: users didn't trust the system to hold what they'd given it. A user saved four items and immediately looked worried. "Does this disappear when I close the tab?" The brand had not yet earned the assumption of reliability. We redesigned around that gap — a persistent counter visible on every page, ownership language instead of utility language. Not "Save for later." "Your collection: 4 items." The difference was not cosmetic. It was the difference between a tool and a relationship.
But Favourites was not a feature. It was brand relationship infrastructure. Every saved item was an intent signal captured upstream — before cart, before checkout, before the brand had traditionally learned anything about what a visitor wanted. For the first time, the brand could see what four million monthly visitors were actually interested in, not just what two percent decided to buy.
The results from the three-market pilot validated the strategy at every level. Conversion increased six percent — twelve million euros annually at the existing revenue base. Feature adoption reached seven percent by month three. Repeat visits among Favourites users rose twenty-three percent. Time between first visit and first purchase shortened by eighteen percent.
The most significant metric was not revenue. It was this: account registrations increased eight percent — not because we had improved the incentive, but because users now had something worth protecting. They had built something. They wanted to secure it. The brand had shifted from asking for commitment to earning it.
The roadmap extended the strategic logic across two additional phases. Notify Me Back converted out-of-stock dead ends into anticipated returns — transforming a brand failure point into a brand promise. Behavioral email activations used intent data to reach users with content drawn from their own expressed interest, not from demographic assumptions. Each phase made the brand more present, more specific, more continuous.
A user in Berlin saved a jacket. Weeks later, she received an automated restock notification. She bought the jacket and added three more pieces she had been considering. "It felt like the site was holding space for me," she wrote. "Like you knew I'd come back."
That sentence became the north star for how the brand talked about digital engagement internally. Not retention. Not re-engagement. Holding space. The brand's job — in every touchpoint, before any transaction — was to make the user feel that their presence had been registered and valued.
The path was not frictionless. Engineering resisted the complexity of anonymous session persistence. Some markets pushed to gate the feature behind login immediately. Senior stakeholders questioned whether deferring purchase decisions would reduce near-term sales. Each objection was answered with the same logic: the feature had to demonstrate value before it asked for data. The brand had to show up before it could ask the user to.
The data made the argument. Users who saved first bought more, not less. The brand's best customers were not impulse buyers. They were people who had taken time to choose.
The work done for Arket changed how H&M Group approaches the relationship between brand strategy and digital architecture. Loyalty is not a marketing program. It is not a discount. It is what happens when a brand behaves consistently with what it says it stands for — in the moments when no transaction is imminent, when the only thing at stake is whether the user decides the brand is worth coming back to.
Arket said: take your time, choose well.
The digital experience now says the same thing.