I'm a builder by nature. I solve problems by writing code, designing systems, and connecting data in ways that make work easier. So when I started noticing a pattern in how our agency was losing clients, my first instinct was to look for the data.
There wasn't any. Not because we didn't care. We cared deeply. But when I tried to understand why a relationship had deteriorated, I'd reconstruct a timeline from Slack threads, email chains, meeting notes in different documents, and whoever had been closest to the account. None of it was systematic. By the time we were doing that reconstruction, we were usually already in damage control.
The analytics blind spot
We had data everywhere. Campaign performance dashboards. Project management tools. CRM entries. Meeting transcripts. The raw material to understand every client relationship in real time was all there. It just wasn't organized into anything useful.
What we lacked was a layer of intelligence that could sit on top of all that data and ask the question we actually needed answered: is this client relationship healthy?
Campaign metrics tell you how the work is performing. Client health tells you whether the relationship is going to survive.
What we actually built
We started building Base for ourselves: an internal system that would ingest meeting data, project status, communication patterns, and client-facing interactions, and surface a real-time view of health across every client relationship we managed.
The early version was rough. But even in its rough form, it changed how we operated. For the first time, we had a single place where a founder or account manager could go before a client call and see: here's the trajectory of this relationship, here's what's changed recently, here's what you should know.
We added Throughline, a conflicts layer that analyzes documents, call transcripts, and messages to identify places where we'd said contradictory things to a client. One of the most consistent patterns in at-risk relationships was small inconsistencies in what different people on our team had communicated. Throughline caught those before they became problems.
The gap is real
When we started talking to other agencies about what we'd built, the response was consistent: they had the same problem and no solution for it. Founders and account managers, a CRM, project management software, but no way to answer which clients are at risk right now.
That's the analytical gap. And it's almost universal in agencies big enough to have complexity, not big enough to have enterprise-grade retention infrastructure.
A systems problem
This isn't a problem you solve with a new process or a better meeting cadence. Those help at the margins. The root issue is structural: the data that would tell you a client relationship is weakening lives in a dozen places and never gets synthesized.
What agencies need is a purpose-built layer that sits across all of those tools, learns from the data they generate, and translates it into something actionable: a health score, an alert, a recommended next step. Building that is exactly what we set out to do.