A relationship book filed as a sales funnel
I had a lead vault problem, not a software problem, and I only saw that after I had built the wrong fix once.
At Common Ground I sit as Chief Operating Officer. The firm is five experts working across capital, brand, technology and operations, and almost everything we do starts with a person we already know. A developer who needs an estimate. A broker with a deal. A sub who bid our jobs last year and will bid the next one. In the spring of 2026 that knowledge lived in four places at once: a large list of raw leads, my phone's contacts, years of message history, and an inbox. None of them talked to each other, and none of them could tell me who we knew at a given company.
On 25 April I moved the book into HubSpot. It worked, technically. The contacts landed, the pipeline was visible, and for a couple of weeks it looked like the job was done.
Then on 6 May I set a standing rule for the firm: every third party we touch goes into the CRM, with contacts, bids, rates and the history of every call and email, and no editorial about whether they were good to work with. The numbers and the dates last; opinions about a vendor go stale in a quarter. That rule is what exposed the problem. A framing sub who bids on our work every season is not a lead moving toward a close. An architect who sends us clients and never buys anything is not a lead at all. HubSpot kept asking which lifecycle stage each of them was in, and the honest answer was that the question did not apply.
What we actually needed was a book shaped like a relationship network. People, companies, and the links between them, queried the way I think about a deal, not the way a marketing team thinks about a lead score.
Move again now, while it is cheap
The easy call was to stay. HubSpot is a good product, it was already paid for, and the records were in. I could have added custom properties, bent the lifecycle stages into something like relationship types, and lived with it. Most firms would.
I chose to move for one reason: the cost of moving only goes up. In mid-May the book was a few thousand records and three weeks of history. A year later it would carry every bid, every call note and every introduction we had logged under the new rule, plus the agents we were starting to build on top of it, all written against the wrong shape. A migration at that point is a project. A migration in the first month is a day.
So on 13 May, inside three weeks of the first move, I moved everything again, this time to a CRM built around records and relationships instead of funnels.
How I came at this one
The question I asked first was what shape the work is, before which tool stands up fastest. That question fit because our book is a network of people and companies that repeats across deals, and a funnel assumes every contact is headed toward one sale. Once the shape was clear, the tool and the timing both followed from it.