Compass's platform was fragmented. CRM, listings, deals, and marketing tools each lived in their own corner of the product, built and shipped as separate efforts over time. Agents filled the gaps between them with spreadsheets and whiteboards, tracking the actual state of their business outside any Compass tool.

The opportunity was to stop treating those tools as separate products and build something workflow-centric instead: a single, centralized workspace organized around how agents actually move a deal from lead to close.

Before designing a single screen, the team mapped the actual buy-side and sell-side workflows agents were running, task by task, from winning a client through post-close. Each stage, prepare the listing, market it, manage offers, close, had its own dense checklist of real work that existing tools didn't account for.

That level of detail mattered. Agent feedback made clear that the product needed to connect the dots between tools, not just centralize them:

That feedback resolved into three concrete needs that shaped everything downstream: agents needed something holistic, a way to understand the state of their entire business and where to prioritize their time; something workflow-based, a way to track everything tied to an active buyer or listing and see what to do next to close it; and something centralized, a single place to collaborate with clients, team members, and everyone else involved in a transaction. Pipeline answered the first need. Folder answered the second and third.

The team synthesized research across four questions, the business problem, the target agent segments, what success would look like, and the benefits agents would actually feel, working through it together as an affinity exercise rather than a single designer's read on the data.

Discovery combined a competitive scan of other real estate tools and deal management systems with direct sessions with agents, sales managers, and internal users. One pattern held across all of it: the high-level stages of a real estate transaction are universal, even though the specific tasks inside each stage vary by team size, agent experience, and market.

That distinction shaped the design approach for the Folder: a modular snapshot view surfaced the information agents needed most at a glance, while still letting them drill into the specific tasks underneath when something needed attention. It's a common pattern in deal management tools, and it gave agents a consistent way to scan a client relationship quickly without losing access to the detail.

I led design across two squads, Pipeline and Folder, and managed two designers within Business Tracker. Compass's design team was organized into several product squads working in parallel, including CRM, an insights team, and a marketing tool that let agents build their own campaigns. Business Tracker's Pipeline and Folder squads sat alongside those, and I owned their design end to end, from early concept through the features that shipped.

Beyond the two squads, I contributed components and guidelines to Compass's growing design system, and ran cross-team sessions, including a workshop I organized and co-facilitated with the marketing product team, to work through how the Folder should represent value from products outside our own squads.

Business Tracker organized an agent's entire book of business around a few core patterns:

The team had about six months to MVP. After initial discovery, research and design ran in parallel with weekly to biweekly feedback sessions, leading into a phased release: a beta with roughly 2,000 agents across different usage cohorts in the fall, followed by a full launch to all agents that January.

Business Tracker saw fast adoption after launch and held onto it.

Business Tracker kept growing after MVP: team collaboration and folder sharing in the first half of the following year, then checklists and role permissioning in the second half. Checklists in particular landed well, becoming the #2 most-requested Business Tracker feature with 280+ votes in Compass's feedback forum, and within a month of shipping, 1,577 agents had discovered it and 452 had added a checklist to a folder.

The six-month MVP timeline was a real constraint. It left no room for deeper IA exploration or rethinking the integrated product workflows the Folder pulled together, so the team leaned on existing Compass design patterns to hit the release date and pushed those improvements to the backlog instead.

Some open questions carried past launch, framed at the time as how-might-we's the team kept coming back to: how to guide agents to actually engage with Compass's other products and features, how to optimize the view so an agent could get an overview of a buyer or listing at a glance, how the information shown should evolve as a deal moved from nurturing through in-contract to close, how to improve consistency in UI and language across a platform the Folder was now consolidating, and how to help agents actually push a transaction forward rather than just observe it.

A RACI chart brought real clarity once ownership was made explicit across teams, something that would have helped earlier given how much of the friction was about who owned what. Content needed to be checked against how agents actually talk, not just internal product language. Leadership expectations around timelines needed active managing, not just a launch date. Looping in product and engineering earlier reduced rework. And having a single source of truth, like the buy-side and sell-side task maps, shared directly with cross-team partners surfaced collaboration opportunities that wouldn't have come up in a design review.