CRM MVP

Roofr was two tools (measurements and proposals) that contractors ran alongside a separate CRM. Every sales call surfaced the same gap: they wanted one place to run jobs from lead to close.

I led the 0-to-1 design end-to-end: problem framing, MVP scoping, interaction model, the high-fidelity prototype used for sales demos, the production spec, and post-launch iteration. The main direction was that the new entity connected existing Roofr entities (measurements, proposals, contacts) under a single new job umbrella.

1. Roofr looked like a disconnected suite, not a platform. Expansion beyond measurement and proposals was capped.
2. Contractors kept a separate CRM open alongside Roofr.
3. Missing foundational CRM blocked the $2B construction-CRM category.

The category positioning itself was at stake. Without a CRM, Roofr was just a high-quality measurement tool. With a CRM, Roofr became the platform for a segment that bloated incumbents (JobNimbus, AccuLynx, JobProgress) were fumbling. The business-model pivot from transactional features to subscription revenue depended on this project clearing the bar.

• Existing entities can’t break. Measurements, proposals, and contacts were already in production with paying customers. The Jobs layer had to connect them, not replace them.
• MVP scope vs 1-year vision. The initial PRD outlined automation, scheduling, and analytics that couldn’t ship in one release. Constant cross-functional re-alignment on what the MVP actually was.
• Performance with large datasets. Eng raised early concerns that Kanban card rendering would degrade at high job counts.
• Sellable before feature-complete. The business needed an sales-demoable story before full implementation. The interactive prototype became a pre-launch sales tool.
• Primary user is an unentrenched contractor. Customers not heavily invested in a competing CRM were the realistic opportunity. Switcher-targeting was a future problem once we had feature parity.

Competitor platforms (JobNimbus, AccuLynx, JobProgress) offered bloated features that overwhelmed smaller ops. That left an opening for a focused, workflow-first MVP rather than matching feature count.

Customers didn’t want new isolated CRM features. They wanted existing Roofr functionality stitched into a cohesive job lifecycle. That reframed the build as a connector layer.

The roofing workflow is high-ambiguity with many edge cases to consider: contacts that aren’t yet jobs, jobs that never get proposals, measurements without jobs attached. The MVP had to support every real state without forcing data work before the first successful capture.

• Scope negotiation is the real work. Cross-functional facilitation to land on “lead to close” mattered more than any screen.
• Architecture beats feature count. Connecting existing entities produced more adoption lift than any new feature would have.
• Prototypes are sales artifacts. The interactive prototype closed sales deals before the build existed.
• Don’t let data cleanliness block capture. Making contacts optional pushed contact engagement up.
• Start the performance conversation early. Eng-side rework on Kanban card rendering could have been smaller with better load testing.

Post-launch. The Jobs layer seeded Custom Workflows, Unified Inbox, automations, suppliers, and more. The foundational architecture we built is the basis for years of future work.