Kimberly Bloomston is the Chief Product Officer of 6sense, a revenue intelligence company.
getty
As product leaders, we've gotten very good at collecting signals. We have not been as deliberate about what happens next.
In many systems, signal collection has effectively become the product. The assumption is that if enough data is captured, the right actions will follow. In practice, that handoff breaks down more often than we’d like to admit. The intelligence is there, but it doesn’t consistently show up in a way that a human can use.
I was reminded of this when I changed jobs recently.
Three vendors reached out. Each one had a signal. None of them had context.
The first got my title wrong. He sent a cold email congratulating me on becoming CEO (I’m CPO) and pitching a product that a CEO would never buy. They had the job change, but they didn’t understand who I was.
The second was a vendor I had worked with across two companies. I had been a champion for them. But then things went south, and we churned them—one of their earliest and largest customers to do so. Sixty days after I joined my current company and became the decision maker again, they scheduled an onboarding call. I had to tell them it was already a churn conversation. They had every signal they needed to know that: the relationship history, the job change, the previous churn. They just couldn’t connect them in a meaningful way.
The third missed a cross-sell opportunity entirely. The signals were there. They just never surfaced at the moment they were needed.
Most business-to-business (B2B) intelligence products are built to collect signals. That's necessary, but it's insufficient. A job change signal tells you something happened, but it doesn't tell you what it means for this account, given this relationship history, at this moment.
The disconnect between those two things is where user trust breaks down.
I've spent considerable time studying why intelligence products lose the confidence of their users. The pattern I keep seeing isn't that the data is wrong, though that does happen. It's that users can't explain what the product is telling them. A seller can't justify a score they don't understand. A marketer can't prove attribution from a signal they can't trace. A CFO asks for evidence, and nobody can produce it. Budget gets cut. The product gets blamed.
But the product is often doing exactly what it was designed to do. It was just designed for the wrong thing.
When a user can't explain what a system is telling them, that's not a user failure. It's a product failure. Specifically, it's a failure to build context into the product's core architecture—to treat signal synthesis and explainability as design requirements rather than features to add later.
Here's where I think most product teams go astray: They treat identity and connectivity as infrastructure problems, and they treat intelligence as a product problem. So they solve them separately, sequentially and at different levels of investment.
The infrastructure team resolves accounts. The product team builds the scoring model. The AI team layers recommendations on top. And somewhere in that handoff, the context gets lost.
What's missing is a deliberate decision to make the connective tissue part of the product. Not just to capture signals, but to reason about them together. Not just to score accounts, but to explain why, in language a seller can use in a conversation. Not just to know that a champion advocated for you at a previous company, but to surface that fact to the team that's about to get on a call with them.
This connective tissue is intelligence put in context. It's not intelligence derived from a single signal, but rather intelligence that emerges from connecting signals across sources, relationships and time.
This requires three things that most products treat as optional:
1. A Coherent Identity Layer: You can't connect signals about a person across companies, roles and behaviors if you don't have a stable, accurate way to identify who that person is. This sounds obvious. It is surprisingly rare in practice. Most companies don't have a consistent customer ID across their own systems, let alone across the external signal sources they're trying to synthesize.
2. Relationship Memory: The signals that matter most aren't always the most recent ones. A former champion who advocated for you two years ago is enormously relevant context when that person just joined a new company. A product that treats every account as a fresh start is missing the most important thing it knows.
3. Explainability By Design: Every recommendation, every score, every prioritization is an implicit claim: "We know something useful about your situation." When that claim can't be explained—when the seller asks "why is this account scoring high?" and the answer is a shrug—the trust breaks. Once trust breaks, usage drops. And once usage drops, the product gets cut, no matter how much value it was theoretically generating.
Before shipping an intelligence feature, I ask: What does the product actually do with what it already knows?
Not what signals it can capture or what the model predicts. Rather, what does it surface, to whom, at what moment and is it in a form they can act on and explain?
The vendors who reached out to me when I changed jobs had signals. They just didn't have a way to make those signals meaningful. The cost of that gap was a missed relationship, a missed retention opportunity and a missed expansion.
That's the cost of treating signal collection as the product. The intelligence was there. The design just didn't surface it.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。