02 — Context

What is the Vendor Rewards Program — and why does it exist?

Delivery Hero operates food delivery platforms in 50+ countries. At the centre of its marketplace model are restaurant and store partners — vendors — who fulfil every order. Their performance directly determines customer experience, platform reliability, and ultimately, revenue.

The Vendor Rewards Program is a tiered performance incentive system designed to motivate these partners to maintain high service quality and achieve operational excellence. Vendors are evaluated monthly across 7 operational metrics and placed into one of three tiers:

Vendors are measured monthly on: orders completed, order rejection rate, offline time, order delay rate, customer contact rate, customer ratings, and menu quality / price parity. Higher tiers unlock better in-app placement, lower commission rates, ad tools, discount vouchers, and top restaurant badges — significant business levers for vendors.

My role & collaboration model

I was the sole designer across the full project lifecycle — from initial problem framing through to global rollout. I defined the UX strategy, conducted vendor research, created wireframes through to high-fidelity production-ready designs, and led design QA. In the final rollout phase, a fellow designer came on board specifically to help carry out region-specific UI tweaks that local market teams had requested. I defined the action items, shared the context, and guided the work throughout — ensuring everything stayed consistent with the system we'd built.

03 — The Problem

A powerful incentive structure — with almost nobody using it.

Before the redesign, the numbers told a story of systemic failure. The rewards program had been built with genuine business intent. But the experience was so confusing, fragmented, and hard to navigate that vendors couldn't engage with it meaningfully — even when they tried.

These weren't just UX problems — they were business problems. The Vendor Rewards Program existed to motivate better operational performance. If 70% of vendors are stuck at Tier 1 and only 7% are clicking through to understand how to improve, the entire incentive mechanism is broken. Higher-performing vendors deliver better customer experiences, generate more orders, and reduce refund costs. Low engagement with rewards meant low improvement of the very metrics DH depended on.

The same information. A fundamentally different experience. — Baseline satisfaction: 3/8 → 7/8 post-redesign

04 — Research & Discovery

Understanding vendors before designing for them.

Before touching any design tool, I needed to understand the real problem from the vendor's perspective. What did they actually see when they opened the rewards page? What were they trying to do? Why were they leaving without acting?

Research followed a continuous iterative loop — not a single discovery phase. I ran multiple rounds of vendor testing and feedback cycles that continued through ideation and prototyping, validating and course-correcting at each stage.

4 core findings from vendor interviews

Every interview surfaced the same four failure modes. The problems were structural — not surface-level.

Competitive benchmarking

I benchmarked rewards and gamification experiences across industries — not just food delivery — to understand what motivates users, how progress is communicated, and where even leading products fall short.

05 — Synthesis & Define

Turning 10 interviews into a shared design direction.

Good research generates noise. Good synthesis generates signal. After clustering findings, three design principles emerged — grounded in vendor psychology, behavioral economics, and scalability constraints — that governed every decision that followed.

Three design principles — grounded in research and design theory

06 — Ideation & Iteration

Four rounds of design. Not one linear process.

The design process was deliberately iterative and user-validated at each step. Not a waterfall — a continuous loop of create, test, learn, and refine. Four distinct rounds, each building on what the previous one revealed.

Iteration is not failure — it's the process. Each round surfaced something the previous one missed.

07 — Design Decisions & Rationale

The calls that shaped the product — and why they were worth making.

Every significant design decision came with a tradeoff. These aren't just "what we designed" — they're the choices where we evaluated alternatives, weighed constraints, and made a call. This is where senior design thinking becomes visible.

08 — Final Solution

What we built — and how it worked.

The redesigned Vendor Rewards hub resolved all four research-identified failure modes: visibility, metric comprehension, benefit discovery, and motivation. Mobile was the primary focus (70% of vendor users) with desktop and tablet fully supported via responsive layouts.

Same component architecture. Different brand expression. Token system in action.

Accessibility & global readiness

The design was built to WCAG-compliant color contrast, typography, and touch targets from day one. RTL support for Arabic and Hebrew was designed in (not retrofitted). Localized date, currency, and number formats were handled through flexible containers. The result: faster launches in 26+ markets with minimal rework per region — a direct reduction in design and development overhead at global scale.

09 — Validation & Experimentation

We didn't just ship it. We proved it.

This is where most portfolios stop: "we tested with users and they liked it." We went further. We built a rigorous, data-first A/B experimentation methodology to generate statistically significant proof that the redesign moved business metrics — not just satisfaction scores.

A/B test structure — powered by Eppo

Targeting configuration: Platform identifiers used to select vendors from specific country instances. Vertical types filtered to restaurants and coffee vendors only — excluding groceries and retail which have different engagement patterns and would have polluted the data signal.

Server-side experiment via Eppo feature-flagging — no performance delays, no UI glitches, and seamless interaction tracking. Each country ran its own controlled test, producing clean, statistically significant results attributable specifically to the design changes.

Metrics tracked

Usability testing results

10 — Impact & Outcomes

The pilot proved it. Then we scaled it.

The redesigned Rewards experience launched first in HungerStation as a controlled pilot. The A/B results gave us the data-backed mandate we needed to scale. Within months, the new design had reached 66% of the total vendor base — with a plan to reach 100% by early 2025.

Reflection

What I learned — and what comes next.

This project taught me as much as it shipped. Four learnings I carry forward as a designer, and five concrete next steps the team has planned.

What comes next — planned next steps