PayPal · Case study
Developer experience, built from scratch.
- Information Architecture
- Design Systems
- Developer Experience
- Content Strategy
- Research Synthesis
- Cross-functional Leadership
- Role
- Lead UX Designer
- Team
- Developer Experiences
- Timeline
- 2020 to 2021
- Platform
- Web · Documentation · Tooling
Overview
A platform left behind.
Developers integrating with PayPal were abandoning onboarding mid-flow, navigating 5 platforms and 24+ screens to reach a working sandbox. The documentation surface, responsible for 80% of all developer site traffic, had no dedicated design function and no shared IA, components, or voice across its properties. Fragmentation wasn’t cosmetic; it was a direct tax on integration velocity.
PayPal’s developer experience had never had a dedicated design function. Documentation, tooling, and onboarding were built independently, each with its own IA, components, and voice.
External research and partner feedback confirmed what the fragmentation showed: the developer surface needed to meet the standard of the consumer product. I came in as the first dedicated design hire on the team, defining scope, running card sorts and tree tests with partners, and owning the IA and system end to end.
Pre-redesign state: three parallel IA models and no shared component library across the documentation surface.
Strategy
Four pillars, one shared standard.
- 01
Unify the documentation
One IA and one design system across 1,700+ developer doc pages.
- 02
Simplify onboarding
Reduce time-to-first-call. The first hour of integration should be a clear path, not a search problem.
- 03
Enhance self-service
Shift volume from support tickets to in-product resolution: stronger search, working examples, real sandboxes.
- 04
Align with modern PayPal
Bring the developer surface up to the quality bar of the PayPal 2.0 brand and consumer product.
Research
Research that shaped the IA.
The program ran a mix of methods against the open questions: internal and external focus groups, card sorts and tree tests against new IA prototypes, and a short developer diary study. Synthesis was iterative, building toward a shared alignment artifact rather than a one-time debrief.
Tree testing on three IA prototypes showed users consistently failed to locate sandbox credentials and integration paths under the legacy taxonomy. That finding drove the shift to use-case-based top-level navigation — the structural decision every subsequent IA iteration was measured against. The output was a developer journey covering discovery, integration, launch, and maintenance that anchored every downstream design decision.
The product team also collaborated on an impact and viability matrix, pairing business objectives against developer pain points to prioritize design sprints. A UX breakdown document translated that matrix into functional priorities and positioning adjustments for the surface.
Design
Search, IA, and navigation
Search rebuilt around query intent
Search was the first surface tackled, and the most broken. A broad audit of PayPal developer docs search mapped query failure patterns and gaps in metadata and tagging. The frames in this rotator are a curated sample of that work, not the full audit inventory. The program migrated to Algolia rather than iterating on the legacy provider, which enabled type-ahead, faceted filtering, and relevance tuning. Shipped improvements leaned on pattern simplification, clearer query results, and stronger page tagging and metadata on the frontend.
Landing page and IA evolution
Eighty percent of all developer site traffic concentrated on the Accept Payments and Checkout area. Which made the landing page and top-level navigation the highest-stakes decisions on the surface. Early iterations tested a product-area hierarchy that mapped PayPal's internal org structure; tree testing showed it failed users trying to locate integration paths by task rather than by product name. The final IA uses a use-case-driven hierarchy, validated across 10+ iterations. The landing page followed: 13 documented states from the original unbranded portal to a PayPal 2.0-aligned final, each tested against the developer journey artifact.
Navigation testing at depth
Validating the new IA required building and testing the models, not just proposing them. Wireframe prototypes were built for each major navigation hypothesis and put through tree tests with internal and external developer cohorts. Over 22 navigation variations were explored before the use-case hierarchy cleared testing. The rejected models fell into two categories: org-mapped structures that required prior knowledge of PayPal's product portfolio, and flat taxonomies that conflated integration guidance with reference documentation.
Documentation systems and sandbox
Custom type ramp for docs scale
The PayPal design system didn't cover documentation density at scale. 1,700+ pages of technical content that needed to be readable, hierarchically clear, and on-brand without the consumer design system team owning the surface. An initial proposal to map developer docs directly to the consumer type scale was rejected: line length and heading density requirements for technical content created legibility failures at the consumer ramp's proportions. The custom scale was researched, pitched, negotiated, and tested against a representative cross-section of page types before approval for rollout.
Simulator: 24 steps to 9
Standing up a test environment previously required navigating 5 platforms and 24+ screens. The new side-by-side simulator pairs GUI controls with live code under a single URL. Pre-issued sandbox credentials removed the account setup detour entirely, cutting the path to a first successful API call from 24 steps to 9.
Payment Simulator: GUI and live code side-by-side under one URL, reducing onboarding from 24 steps across 5 platforms to 9.
Outcome
Four outcomes, one coherent surface.
88+
page templates designed and tested
10+
navigation IA iterations tested
24 → 9
steps to a working sandbox
1,700+
docs unified under a new IA and design system
Reflection
Time, trust, and findability.
For developers, time is the constraint. They need to assess a product quickly and commit to an integration path without losing hours to navigation. Cohesive style and continuity across PayPal’s landing, docs, search, and sandbox made that first read trustworthy and kept them moving toward the workflow they came for.
Repeated IA redesign and tree testing were aimed at that clock. Each round asked whether someone could reach the right integration decision quickly and accurately, not whether the site merely looked reorganized.
Previously
No earlier case study yet
Next case study
Under Armour Unified Profile →