Rebuilding RevOps for a scaling DevOps company

The engagement

This is our deepest ongoing engagement: a weekly delivery cadence with the company’s RevOps director, who defines priorities while we design and build. Over multiple quarters the work has covered nearly every part of the revenue engine. Here are the pieces that mattered most.

What we built

A scoring model the whole team trusts

Marketing and sales had no shared definition of a qualified lead. We designed and shipped a two-tier MQL scoring model: physical events score differently for a badge scan versus an actual conversation, virtual events score differently for registering versus attending, and a booked demo meeting triggers MQL instantly. We found and worked around a real HubSpot limitation along the way: later low-value actions can overwrite higher scores, so we anchored the model with static lists.

Account tiering that took one week, not one quarter

The first tiering attempt was a complex automation. We killed it and did the simple thing: tier every account against a verified customer list, add industry and spend-size properties, and bulk-reassign. 2,207 accounts got a tier, and a monthly QA report now catches untagged new companies.

Pipeline architecture built for ARR

New business, renewals, and expansion were mixed together, so forecasting was guesswork. We separated them: a Sales pipeline for new business and expansion, and a dedicated Renewals pipeline with time-based stages and automated notifications starting 90 days out. ARR and TCV are now standardized on line items, multi-year deals are modeled year by year, and expansion deals can’t close without linking to their renewal. Win rate from first meeting became a measurable number: 18.5%.

Closed-lost deals that come back

Lost deals used to disappear. Now a staged task sequence re-engages them: three touches, two business days apart, assigned to the current account owner with a per-rep priority queue. Leadership runs a formal closed-lost review on top of it.

Buyer intent, routed to the people who sell

We connected HubSpot to the intent platform two ways: intent signals (research and visitor activity) flow in for top-tier accounts, and the signals surface directly in the AE dashboard rather than another notification channel nobody checks.

AI call intelligence

The team is migrating call recording to Zoom, with an external Claude workflow that analyzes transcripts for manager summaries and call quality review. Meeting types are standardized so each kind of call gets the right analysis.

What this looks like day to day

No six-month roadmap document. A weekly meeting sets priorities, we ship during the week, and finished work appears in the next meeting’s agenda as “live.” In a typical week that’s 10 to 15 shipped items: workflows, dashboards, imports, property architecture, and fixes nobody else has time for.