B2C · Marketplace · In-House

HeyCar Growth Experiments & Mobile Platforms

Combined rapid experimentation with a native mobile bet so HeyCar could close the gap to mobile first marketplaces, double engagement where measured, and ship leadership aligned lead generation.

Role

Experience Design & Labs Lead

Context

HeyCar (VW Group) · Berlin

Period

2020 to 2021

Platform

iOS · Android

Executive Summary

HeyCar entered Germany’s used car marketplaces behind Mobile.de and AutoScout24 on native quality while holding strong web traffic that did not convert intent at parity.

I led research and design for growth experiments and HeyCar’s first native apps, aligning executives on a mobile first roadmap and shared funnel instrumentation.

Engagement roughly doubled on measured journeys, native apps beat web on lead generation, and leadership adopted experimentation as a repeatable portfolio muscle.

The core mission

Ship evidence led bets: validate costly service ideas before logistics spend, then win the native surface where German shoppers already compare inventory, price, and dealer trust on their phones.

Strategic Landscape

Germany’s classified leaders already trained buyers to expect fast filters, saved searches, and dealer outreach from a pocket sized product. HeyCar’s responsive web footprint captured curiosity, yet the strategic risk was existential in a category where switching costs are low and inventory discovery is the product.

Volkswagen Group backing gave us brand trust and supply side access, but not infinite runway to chase every service idea. If we stayed web only while competitors doubled down on native retention loops, we would keep paying acquisition for traffic that leaked before high intent leads reached dealers.

Diagnosis

The gap was not “we need more features.” It was a portfolio question on where to concentrate design, engineering, and data science so the next euro of investment improved measurable intent, not vanity traffic.

User problem

Shoppers researched on phones while standing in parking lots and on commutes. Mobile web introduced friction just as they compared stock, financing hooks, and dealer response times. Competitors’ native apps set the baseline for speed, saved flows, and perceived reliability.

Business problem

Lead volume without intent quality wastes dealer capacity and erodes trust in the marketplace. Leadership needed to decide whether native investment was mandatory, and which adjacent services deserved MVP level funding versus a cheaper proof.

Organizational problem

Strategic design, growth, product, and data science each held pieces of insight, but no single squad owned “mobile marketplace outcomes” end to end. Without shared instrumentation, teams debated opinions instead of converging on funnel metrics tied to dealer revenue.

Team and operating model

Experiments only work when design, PM, engineering, and data science share a hypothesis sheet and the same funnel IDs. I stayed embedded with our data partners so research stimuli, feature flags, and success criteria stayed aligned from sketch to readout.

Workshop notes and experiment planning artifacts on a wall — Experimentation culture: validate ideas through rapid prototyping at scale, build fast, stay metrics driven, and keep data science in the loop from the start.

Strategic Response

We would not greenlight logistics heavy services without behavioral proof, and we would not debate native forever while competitors widened the retention gap.

Instrument before you argue

We paired qualitative signals with warehouse scale analytics so every experiment wrote directly to leadership ready KPIs, not slide filler.

Fake door over fleet

For premium at home test drives, we shipped an in product funnel with paid and free variants instead of building dispatch ops first. Intent beats opinion.

Prestige targeting via data science

Strategic design’s Prestige persona became operational: data science segmented returning users so we could target high propensity cohorts without spamming the whole base.

Native as the lead KPI surface

Once mobile web metrics showed the engagement ceiling, we reframed native apps as the lead generation product, not a marketing wrapper.

Split design sprint ownership

Exec heavy calendars could not sit five consecutive days. Core leaders locked the storyboard in three days, heyLabs executed build and usability passes without losing narrative continuity.

Trade-off: We sacrificed a textbook five day sprint and some polish velocity in exchange for leadership attendance, faster learning cycles, and a single accountable squad narrative. Interest and intent lived in one funnel (postal code for curiosity, email for sales ready follow up) while lead generation stayed the executive north star for the native launch.

From signals to shipped bets

Each milestone moved us from abstract strategy to something executives could screenshot.

Market signal

Industry research plus our own behavioral data showed test drives remained a conversion choke point worth exploring before we scaled listings alone.

Hack week commitment

Quarterly innovation time became the forcing function: one week, one hypothesis, one readout. No anonymous side quests.

Client side experimentation

We extended feature flags with userID based assignment so designers could ship variants without waiting for a full server rewrite.

Executive readout

When funnel metrics moved, we brought managing directors the same artifacts the squad used, not a summarized version. Transparency accelerated the native mandate.

Experiment funnel variants showing paid and free booking paths — How we differentiated the two offers: Variant A, free to book cars; Variant B, 100 EUR to book up to three cars, to compare behavior across segments.

Design & Delivery

heyChauffeur: validate demand before wheels turn

Research finding

Consumers wanted flexible test drives, but logistics, insurance, and depreciation made a full MVP ruinously expensive if the hypothesis failed.

Design decision

We prototyped an in platform booking funnel with two offers (free versus 100 EUR for up to three cars) and measured postal code plus email steps as separate intent layers.

Mobile UI for the heyChauffeur experiment — User flow for heyChauffeur: postal code captures interest, email captures intent, so we could validate demand on the platform before committing to real world service logistics.

Research finding

We needed Prestige segment truth without waiting for months of qual recruitment.

Design decision

Data science targeted Prestige userIDs on return visits, while we mirrored the same UX for broader cohorts to compare behavior, not just anecdotes.

Native apps: sprint, learn, ship

Research finding

Mobile web attracted phone traffic, but usability studies showed persistent friction in filters, onboarding, and dealer handoff compared to native patterns users already knew from competitors.

Design decision

We ran a condensed design sprint with CPO and strategic design in the room, locked a storyboard, then let heyLabs iterate prototypes against lead generation KPIs before App Store and Play submission.

Design sprint workshop with team reviewing storyboards — The last day of the design sprint as the team decides on the scope of the final prototype.

HeyCar mobile app interface — Shipped native apps on the App Store and Google Play, with flows refined toward lead generation as the primary KPI: prospective customers contacting dealerships through the platform.

Impact

2×

Engagement uplift on instrumented mobile journeys after we prioritized native patterns and experimentation learnings in the roadmap.

Lead gen

Native apps beat the web product on lead generation, the KPI leadership used to justify continued investment in mobile.

Validated

heyChauffeur funnel proved measurable intent before we committed to logistics, insurance, and fleet coordination spend.

These outcomes were reviewed with managing directors using the same dashboards the squad watched daily, not a sanitized monthly rollup.

Transformation

Beyond the metrics, HeyCar learned to treat growth as a portfolio of bets with kill criteria, not a sequence of heroic launches.

→

Evidence led roadmap

Opinion driven debates gave way to experiment readouts that fit executive calendars. Mobile first stopped being a slogan and became the planning default.

→

Shared tooling

Client side experimentation patterns became reusable infrastructure other teams could adopt without reinventing governance.

→

Design at the center

Design connected dealer economics, consumer psychology, and technical constraints so narratives stayed coherent from hack week slide to App Store screenshot.

Leadership Principles

Prove intent before ops

The cheapest learning happens in the funnel, not in the warehouse. If you cannot sketch the KPI sheet first, you are funding theater.

Adapt rituals to power calendars

A perfect five day sprint that no executive attends delivers slower truth than an adapted sprint with explicit handoffs.

Make data science a co author

When designers own the hypothesis and data science owns the cohort keys, you avoid beautiful interfaces that cannot be targeted or measured.

Looking ahead

The native apps gave HeyCar a credible pocket presence, while the experimentation stack created room for more service bets at bounded risk. Next horizons were deeper personalization on listings, dealer side workflow parity, and extending validated services into fulfillment only when metrics repeated in multiple regions.