Scaling First-Party Advertising Through Self-Serve Design
How we designed a self-serve advertising platform that enabled Kleinanzeigen to reduce operational overhead, empower advertisers directly, and build the foundation for a retail media business.
From manual operations to scalable retail media
Kleinanzeigen operates both third-party and first-party advertising. To scale revenue and reduce operational overhead, we designed and launched a self-serve platform enabling advertisers to create and manage first-party campaigns independently.
My role was to co-own the design strategy and execution across the entire initiative — from prioritization frameworks and MVP definition to shipping core features and validating post-launch iterations.
A structural shift in advertising strategy
Kleinanzeigen historically relied on third-party advertising networks. But structural risks and market trends were forcing a strategic pivot.
Platform dependency
Heavy reliance on Big Tech ad networks (e.g. AdSense deprecation risks) and third-party cookies amid tightening privacy regulations.
Operational bottleneck
All first-party campaigns were managed manually by Customer Success — limiting scalability, slowing onboarding, and increasing cost.
Retail media opportunity
Marketplaces like Amazon and Mercado Libre proved that first-party data and high purchase-intent environments generate substantial advertiser value.
“If Kleinanzeigen wanted to scale first-party advertising to 50% of revenue, the manual model would not hold. A self-serve platform was necessary.”
Not just a dashboard — a new advertising infrastructure
Translating strategic ambition into a product surfaced a much more complex challenge across four layers.
Scalability
The current model required internal teams to upload feeds, adjust budgets, and interpret performance. Scaling revenue would have meant scaling headcount proportionally. The product needed to remove this bottleneck entirely.
Segmentation
Advertisers were not homogeneous. Network providers needed automation and API access; retailers needed control and dashboard interaction. A single one-size-fits-all solution would fail both segments.
Performance expectations
The shift from visibility ads to performance-driven retail media meant advertisers needed measurable ROI. The UX had to support decision-making, not just configuration.
Marketplace integrity
Monetization could not erode user trust. Poor ad placement or irrelevant promotion would increase churn and undermine the core value proposition. Relevance and contextual targeting had to be built in.
“How might we design a scalable self-serve advertising platform that empowers both automation-driven network providers and hands-on retailers — while replacing manual workflows and protecting marketplace user experience?”
Reducing opinion-driven decisions before building
Before defining the MVP, we had a long wishlist, market expectations shaped by large retail media players, and strategic pressure to deliver fast. We structured discovery around three complementary inputs.
KANO Survey — 7 strategic advertising clients
Instead of asking “Do you like this feature?” we asked how users would feel if a feature existed — and if it didn't. This let us separate what drives satisfaction from what's merely expected.
- Campaign creation & configuration
- Budget management
- Basic performance reporting
- Campaign recommendations & alerts
- Positive & negative keyword targeting
- Auto-targeting with manual override
- Smart bidding with max CPC guardrails
- Budget utilization heatmaps
- Feed management automation
- Platform & geo targeting
- Individual ad assignment
- Campaign reach forecasting
Segmentation — Two fundamentally different advertiser archetypes
Network Providers
- Large, dynamic product feeds
- Preference for automation & APIs
- ROI and reach-oriented
- Managed-service mindset
Retailers
- Stable product inventories
- Desire for control & visibility
- Interest in audience targeting
- Comfortable with dashboards
Instead of designing two separate products, we decided to design a single flexible system capable of adapting to both — a decision that shaped the entire product architecture.
The smallest viable platform that delivers real value to both segments
I co-led a cross-functional MVP scoping workshop with UX Research to align on scope, evaluate architecture directions, and translate research into structural product decisions.
✅ Included in MVP
- → Feed ingestion & management
- → Campaign creation & budget configuration
- → Keyword targeting (positive & negative)
- → Automated bidding with max CPC guardrails
- → Reporting dashboards & performance insights
- → Alerts & recommendations
⏸ Intentionally deferred
- — Platform & geo targeting
- — Advanced reach forecasting
- — Overly granular ad assignment
- — Individual campaign reach metrics
Resolving the automation vs. control tension
Rather than creating two different systems, we designed one flexible framework built on a single guiding principle.
“Automation by default. Manual control by choice.”
Building confidence before control
The home dashboard had one core job: give advertisers immediate confidence in performance without forcing action. Automation users needed reassurance (“Is the system working?”). Control-oriented users needed context (“What exactly is happening?”).
We prioritized a clear KPI hierarchy — Daily Budget, Spend, Impressions, Clicks, CTR, Average CPC — with WoW comparisons and progressive entry points into deeper sections.
Turning data into decisions
The Performance section is where control-oriented advertisers live. The challenge was providing depth without overwhelming smaller clients. We structured the page into clear drill-down layers: summary → trend comparison → filtered view → budget breakdown → campaign table.
A per-hour budget heatmap made pacing visible without requiring manual configuration — supporting the “automation with guardrails” principle for both segments.
Where automation and control intersect
The campaign detail page is the core of the system. CPC and daily budget are editable, but bidding is structured. Max CPC acts as a safety boundary — preventing over-optimization and budget waste, while preserving advertiser agency.
A progressive editing model meant retailers could intervene directly, while network providers experienced minimal friction. Core parameters were editable; advanced mechanics were abstracted behind progressive disclosure.
Automation by default
Smart bidding and targeting operate in the background — reducing friction and cognitive load for all users.
Manual override
Budgets, CPC boundaries, and campaign parameters are always editable — control is never hidden, just not mandatory.
Progressive disclosure
Basic users see clarity. Advanced users can dig deeper. Complexity is accessible but never imposed.
The Advertising Detail Page — improving click quality without sacrificing revenue
After launching the platform, we identified a structural risk in the monetization model: the immediate redirect on click generated traffic, but not all of it was high-intent. Advertisers were scrutinizing conversion quality, bounce rates, and ROAS — and the numbers told a concerning story.
The core issues were clear: low click quality, no engagement layer before exit, inconsistent UX transition, and no opportunity to leverage trust signals or similar ads modules.
Problems with the current model
- ✕ Immediate redirect — no intent confirmation
- ✕ No contextual reinforcement before leaving KA
- ✕ Limited engagement tracking
- ✕ High accidental / low-intent traffic
- ✕ No room for similar ads or trust signals
“How might we increase click quality and engagement without breaking the current CPC model, creating excessive friction, or damaging short-term revenue?”
The solution: an intermediate branded page within Kleinanzeigen
Increase intentionality
The first click becomes a billable event. The second click — “Continue to Advertiser” — confirms real intent. This transforms CPC from a pure click metric into a richer funnel model with ADP impressions, click-outs, and bounces.
Preserve brand continuity
The ADP displays product images, price, seller badges, category breadcrumbs, and a “Sponsored” label — creating visual continuity and trust reinforcement before the user exits to the advertiser's site.
Enable future monetization
The ADP opens the door for similar ads modules, premium placements, brand blocks, and engagement-based pricing models. It becomes infrastructure, not just a page.
Validating the business case before full rollout
The ADP introduced a real business tension: an extra step could reduce outbound traffic and temporarily lower CPC revenue. We ran a phased A/B test to measure the trade-off between short-term revenue risk and long-term advertiser value.
Direct redirect
Users clicking on a sponsored ad are immediately redirected to the advertiser's external website. Standard CPC model with no intermediate page.
Advertising Detail Page
Users land on an intermediate branded page within Kleinanzeigen. They see product details and must explicitly confirm intent before exiting via “Continue to Advertiser.”
Using AI to accelerate exploration, not replace thinking
With multiple advertiser archetypes, a tight Q3 deadline, and constant cross-team alignment pressure, traditional design cycles would have slowed us down. We integrated AI and Figma Make into our early-stage exploration to reduce friction — not to replace product thinking.
AI helped us generate structural layout variations (dashboard hierarchies, automation toggle patterns, campaign configuration models) and compare mental models faster. Figma Make then converted ideas into clickable flows, shifting stakeholder conversations from “what do you mean?” to “is this the right model?” — significantly accelerating alignment.
AI did not define strategy, replace research, or decide final UX patterns. Human synthesis remained the critical layer throughout.
Platform shipped. Infrastructure built. Iteration validated.
Q3 2025 Launch
The self-serve platform launched on schedule, replacing fully manual Customer Success workflows. Advertisers could now create and manage first-party campaigns independently.
Retail media foundation
The platform established the core infrastructure — feed management, smart bidding, keyword targeting, reporting — needed to scale first-party advertising toward the 50% revenue target.
ADP A/B test initiated
Post-launch, we designed and ran a structured A/B test to validate the Advertising Detail Page concept and determine its impact on click quality and advertiser value before full rollout.
Key takeaways
Segmentation is a design constraint, not just a marketing concept
Identifying two fundamentally different advertiser archetypes early wasn't a research exercise — it materially shaped every architecture decision, from feed management to reporting depth. Skipping this would have led to a product that served neither group well.
Scope clarity is a design deliverable
Co-leading the MVP workshop was as valuable as any screen we designed. Translating research into scope decisions — and explicitly deciding what not to build — gave the team a shared foundation that reduced rework throughout the project.
Business tension makes for better product thinking
The ADP was genuinely difficult to champion because it introduced short-term revenue risk. Designing for that tension — and proposing a structured A/B test as the path forward — was more valuable than a clean solution that ignored the constraint.