- Design systems
- Workflow automation
- Partner UX
Partner onboarding at scale
Problem
TiVo’s partner ingest process — the way content providers got their catalogs into the TiVo metadata system — had grown into a tangle of email threads, spreadsheets, and one-off engineering escalations. Each new partner cost the team weeks of manual coordination, and the partners themselves had no shared view of where their batch stood.
The cost wasn’t only in engineering time. The slow onboarding was a ceiling on how many partners the metadata business could absorb in a quarter, and it pushed the work toward the partners who already knew us well enough to navigate the process.
Approach
I led the visual and interaction design for a single, structured partner workflow — anchored on a partner-facing portal and an internal review surface that shared the same underlying state.
Discovery
The team had been treating onboarding as an engineering throughput problem. Partner conversations surfaced something different: the friction wasn’t the work itself, it was the lack of a shared definition of “done” for a content batch, and the absence of a single place where status was authoritative. The redesign followed that insight.
Information architecture
I modeled onboarding as a small set of named phases — each with its own surface, its own gates, and its own clear next action. Partners always knew what they owed us; we always knew what we owed them.
Component system
The portal needed to look and feel like one product even though it wired into multiple internal systems. I built a small design-system layer — surface tokens, status pills, a unified upload pattern, a validation table — that could absorb the back-end heterogeneity without leaking into the partner experience.
Validation surface
Validation feedback was where the old process broke down. The redesign moves validation inside the workflow — errors inline, errors linkable, a clear path from “what’s wrong” to “how do I fix it” without a support ticket.
Outcome
Reflection
The most useful design decision wasn’t the visual system or the upload flow — it was insisting that the partner and the internal reviewer were looking at the same surface, with the same status, at the same time. Most of the original pain was the cost of two parties disagreeing about what state a batch was in. Once that collapsed into one shared view, the rest of the workflow followed.
There’s a version of this project that gets sold internally as “we shipped a portal.” The version I keep returning to is: we shipped a shared mental model.