2024 Custom Workflows
Turned a single static pipeline into three customizable workflows, so roofers can model insurance, residential, and repair jobs the way they actually operate.
Case highlights
3,070
Custom stages created
across the customer base, 1 year post-launch
37%
Elite users on all 3 workflows
another 25% on 2 workflows; rest on 1
99%
Coverage by 3-workflow cap
~30 customers requested a 4th; backend supports unlimited
The CRM MVP shipped with a single static pipeline. It worked for residential only contractors, but made it difficult for those that quoted on insurance and repair jobs, because they have different lifecycles. I led the design end-to-end, discovery, the entity model with eng, settings surface, multi-workflow Board View, tier-gating, and post-launch analytics. To avoid the complexity of going unlimited workflows we went with three per company, each customizable and sharing a stage pool so the mental model stayed simple.
- The single MVP pipeline cut out entire customer segments. Insurance and repair contractors felt frustrated with the limitation.
- Stages weren’t customizable. Contractors wanted basic functionality like custom naming.
- Customers running mixed operations (60% residential, 40% insurance) were forced into one lifecycle for both.
Customization was the second-loudest product request after the CRM itself, because contractors wanted the board to model their actual workflow.
Launch
Adoption signals
- 01 Adoption of more than 1 workflow per company.
- 02 Customization of stages.
- 03 Customer feedback confirms the segment gap closed.
Strategy
Operating-system claim
- 01 Insurance + repair segments unblocked.
- 02 Elite-tier upgrade lever.
- 03 Settings page remains easy to understand.
- 04 Architecture leaves room to unlock more workflows later.
- Can’t break existing jobs. Old behavior unchanged, new customization opt-in.
- Default path stays fast. Most customers won’t customize; settings can’t be a tax on the 80%.
- Artificial cap. Stick to 3 but need the option to raise in the future.
- Tier gated. Had to feel like it was actually adding value.
The hypothesis to disprove was “customers want a blank workflow builder.” Interviews killed it fast. Most contractors wanted tweaks (rename, add “Follow Up”, delete an unused stage).
Competitor teardown (JobNimbus, AccuLynx) confirmed the trap: deep customization, rarely used. Their UI added significant complexity that contractors didn’t need or want.
Most customers run two or three processes that share most stages, with only a handful of segment-specific ones (Claim Filed, Adjuster Meeting). That overlap seemed to suggest a shared-stage pool would work best.
Single static pipeline, on/off list of stages. Worked for residential, frustrated everyone else.
Two ideas surfaced early: starter templates instead of a blank canvas and a shared stage pool so a stage named once could appear in multiple workflows without a confusing data model.
Settings surface and the multi-workflow Board View, both shipped to Elite. Hard cap of 3 workflows; backend supports unlimited.
Workflow management. Up to 3 workflows, Manage per row, cap-reached state at 3.
Stage editor inside a workflow. Drag-reorder, add-from-pool, rename, disable.
REEL Multi-workflow Board View
All workflows by default. Workflow chips on every card, optional single-workflow filter. The unified board is what makes customization feel like a platform capability, not a settings page feature.
Shared stage pool
Stages live at the company level and are referenced by workflows.
Suggested workflows
Three pre-built workflows instead of a blank builder. Customers start from a template and tweak.
Soft cap of 3 workflows
Backend supports unlimited; UI caps at 3 to cover most customers without bloating the surface. ~30 requests for a 4th post-launch.
Elite-tier gating
Reinforced the subscription pivot. Slowed adoption among non-Elite tiers by design; lifted Elite value.
Unified Board View
All workflows on one board with chips per card. Optional single-workflow filter.
-
3,000+
Custom stages created
1 year post-launch. Top added customs: Scheduled, Claim Filed, Adjuster Meeting, Approved, Prospect, Follow Up, Gutters. Mostly insurance and repair vocabulary, exactly the use case customization was scoped for.
-
37%
Elite users on all 3 workflows
Another ~25% on 2 workflows and the rest stayed on 1.
-
99%
Coverage by 3-workflow cap
~30 customers requested a 4th. Backend supports unlimited; the cap is a product decision driven by settings-surface usability, not a technical constraint.
-
Proposal Signed
Most-renamed shared stage
Once renamed on one workflow, the shared-stage architecture propagated across all three.
-
Operating system
Positioning made real
Insurance and repair now have a working lifecycle inside Roofr.
Tier-gate caveat. Adoption is measured against the Elite-tier population, the only tier with workflow-customization access.
- Disprove the customization assumption early. “Customers want a blank builder” was wrong in workflow-heavy B2B. Final scope was much smaller than the PRD implied.
- Architecture decisions shape the UX. The shared-stage pool is invisible but makes every later decision easier.
- Tier gates make usage data harder to read. 37% adoption is 37% of a filtered population, not the full demand curve.
- If repeating: the multi-workflow Board View rendering was underscoped. Treat the view layer as its own scoping exercise next time.
Post-launch. Custom Workflows is the follow-up to the CRM MVP that helped expand the prospective audience.
Deep dive
Curious about the specifics?
Adoption cohorts, revenue pivot data, prototype walkthrough videos, and the internal PRD are available on request during a portfolio review.