Making Lounge personally relevant.

Summary

Lounge by Zalando is Europe’s largest flash-sale fashion platform: 2.5M monthly users across 20+ markets, campaigns that expire in 2 to 3 days. Users faced 35 to 45 campaigns a day, 1,000+ products each, ordered by recency. Position was the only factor driving clicks, so more than half landed on the top five. The catalog was set to scale past 100 campaigns a day. At that point this stopped being a UX problem and became a discovery gap with direct impact on conversion and GMV.

The business assumed personal relevance wasn’t realistic at Lounge. I used the 2019 redesign of the campaign view to prove otherwise. Personal relevance outgrew the project into a dedicated workstream, then became a core opportunity area of Lounge’s 2023 product vision.

01 / The Challenge

A homepage about to break under scale

My Lounge was the page every Lounge customer landed on after login: 35 to 45 live campaigns a day, sorted by freshness.

Position drove clicks, almost on its own. Campaigns near the top got the engagement, almost regardless of the content. With the business planning to scale the campaign volume significantly, relevant campaigns would only get buried deeper.

51%

of clicks landed on the top five campaigns

34–45

campaigns per day, sorted by freshness

80%

of traffic on mobile

In 2018 a redesign of the campaign view had failed. Several A/B tests ran where the redesigned components performed significantly worse than the control.

02 / Diagnosis

The brief was layout; the real problem was relevance.

The challenge was not just helping users navigate a large catalog quickly, it was making sure each customer saw first what mattered to them. Personalised experiences had become an established pattern in the industry, a capability Zalando had already validated on its core app.

03 / The Reframe

Pursuing personal relevance

The sprint set out to help customers find relevant campaigns faster. Structural changes would help, but position drove clicks almost on its own, so the bigger lever was relevance: the right campaigns at the top, for everyone.

We interviewed experts from the team owning customer preference data at Zalando main store, and people at Lounge working on scattered initiatives trying to build relevance.

We built a separate high-fidelity prototype for each participant, using their real browsing data. To test the value of a personalised experience, it had to feel real.

04 / The Proof

The data made customisation redundant

The work on the redesign led to a dedicated personalisation workstream at Lounge, which ran the first brand relevance experiment on the campaign overview, up-sorting campaigns based on the brands a customer already preferred.

01 · Positive results

A statistically significant lift in Orders per User and a measurable drop in Exit Rate. The uplifts were small, but they proved the direction was right.

02 · Learning

Brand was central to why people came to Lounge, but brand alone was not a strong enough signal to make the experience feel personal.

03 · Decision

Customisation feature was dropped. It was the right move when we could not read behavioural data, and the wrong one once the partnership gave us real preference data.

05 / The Strategy

The workstream becomes an opportunity area

In December 2020, Lounge launched its 2023 Strategy: eight opportunity areas to define the product vision for the next three years. The PM I’d worked with on the campaign view redesign and I became the strategy drivers for Personal Relevance.

8.6%

12-month retention: 3.8M active of 44M registered

~4 / 55

campaigns engaged per session, in 6 to 7 minutes

300/wk

campaign volume the business was scaling toward

The arithmetic was the case for the work. Stock was going broader and shallower, more lines, fewer units each. Campaign volume was climbing toward 300 a week. The same uniform list shown to every customer was becoming impossible to sustain, and the customers it served worst were the ones already leaving.

06 / The Constraint

Off-price personalisation is a different problem

Personalisation on the parent Zalando platform does not transfer to Lounge. Zalando works with deep, replenishable stock, so it can optimise around the customer alone. Lounge cannot: its stock is finite and its campaigns are short, only a few days each, so relevance has to help sell the stock that is live, not only suit the customer. That changes the question the model has to answer.

07 / The Frame

The three problems the team aligned on

I synthesised over ten prior research efforts, qualitative and quantitative, NPS and onsite feedback, into a single body of evidence that fed the problem-framing workshop. From it, we defined three clear, non-overlapping problems, each framed as a How Might We statement.

The parent Zalando platform’s research team had already built a six-portrait segmentation. Rather than author our own, I selected the three that fit Lounge’s base and applied them.

08 / The Vision

From concepts to C-Level approval

Rather than solving the three problems within today’s constraints, we used them as the starting point to imagine the experience three years ahead, unconstrained and deliberately optimistic, then walked backwards to what could be built. Adapting Airbnb’s eleven-star method to a strategy problem, I framed four concepts as ten-star experience narratives. Not features. Pictures of what relevance could feel like at its most ambitious.

A roadmap aligns a team for a quarter; a vision aligns a company for three years. Each narrative was rooted in a documented insight and one of the three customer portraits.

Seven editorial revisions and two C-Level approval cycles later, the vision and its concepts passed. I handed the Driver role to a Principal Designer, and the product groups took the concepts into delivery.

Coda

What shipped, and what it returned

The first feature built on the approved direction was Top Picks for You: a single curated campaign aggregating the most relevant products from across all live campaigns, the concept I had drafted as Curated Just for You.