Growth-stage companies repeat the same decisions because no single tool in their stack sees the whole picture. The problem isn't missing data. It's a missing synthesis layer that connects what each tool sees separately.

Why do growth-stage companies repeat the same decisions?

The mechanism is always the same. A team sees a signal in one part of their stack — acquisition numbers look strong, a feature shows low engagement, a customer segment looks ripe for expansion — and makes a decision based on it. The decision feels data-driven because it is, in a narrow sense. The problem is that the signal came from one tool, and no one asked what the other tools were seeing at the same time.

Six months later, a different team faces a related question. They pull from their tool. They see their slice of the picture. They make a similar call, or reverse the previous one, without knowing the earlier decision was made or why. The loop closes. Teams cover the same ground twice.

This isn't a communication failure. It's a structural one. There's no layer in the stack whose job is to connect what each tool sees and surface the full picture before anyone makes a call. Each team operates from their corner, and the same blind spots recur.

What are the three layers of a growth stack?

To understand what's missing, it helps to name what most growth stacks actually contain.

Three Layers of a Growth Stack

Layer 1 - Data collection

The tools: Mixpanel, Amplitude, Google Ads, Salesforce, Intercom, Klaviyo, and the rest. Each one captures a stream of behavior, signals, or outcomes in its domain. This layer is well-developed at most growth-stage companies. The data exists.

Layer 2 - Reporting

Dashboards, BI tools, weekly metrics reviews. This layer takes the data from Layer 1 and shows you what happened — in one tool, over one time window, for one team's questions. Looker, Tableau, native analytics dashboards. Most companies invest heavily here. The reports exist.

Layer 3 - Synthesis

This is the missing layer. It doesn't show you what happened in one tool. It connects what multiple tools saw and tells you what it means together. It surfaces patterns you didn't know to look for, because it's comparing signals across systems that were never designed to talk to each other.

Most growth-stage companies have Layer 1 and Layer 2. Almost none have Layer 3.

The assumption is that Layer 2 covers the gap — if you build enough dashboards, read them carefully enough, and get your teams to share their numbers in the same meeting, the synthesis will happen. It doesn't.

What is missing from a typical growth stack?

The synthesis layer. And the distinction matters because it changes what you build next.

If the problem were a Layer 1 problem, the solution would be another data source. If it were a Layer 2 problem, the solution would be better dashboards or more rigorous reporting. When the problem is Layer 3, adding more tools or more dashboards makes it worse. You're adding to Layers 1 and 2, which already have enough data. The missing layer requires a different kind of function.

Growth teams often sense this without naming it. The instinct shows up as: "We have all this data and still can't get a clear answer." The data team is swamped building reports that teams don't use. Dashboards proliferate. Meetings get longer. The signal-to-noise ratio drops.

The missing layer is a synthesis function — one whose job is to sit above the tools and ask what they're telling you together. In a mature company, this is what a strong insights or research function does. At a growth-stage company, it rarely exists as a dedicated function. It exists, if at all, as a talented analyst/product manager/marketer who triangulates informally and is usually stretched too thin to do it consistently.

How is an intelligence layer different from a dashboard?

A dashboard shows you what happened in one system. An intelligence layer connects what multiple systems saw and tells you what it means.

A dashboard answers questions you already knew to ask. You decide what to measure, build the chart, read the number. Reactive by design — input in, output out.

An intelligence layer surfaces patterns you didn't know to look for. It compares your acquisition data against your activation data, your support queue against your NPS curve, your sales signals against your product usage. It finds the contradictions and the correlations that no single tool can see on its own.

Two Failures of Fragmented Data — the Invisible Cohort and the Wrong Signal — both happen in Layer 2. The dashboards work perfectly. The reports are accurate. Without a Layer 3 connecting them, the contradictions stay invisible until something breaks.

A dashboard is a window into one room. An intelligence layer is the view from above the whole building.

In the next post, I'll get into what the synthesis layer actually does when it's working — specifically, the difference between intelligence that arrives only when you ask for it and intelligence that surfaces before you know to look.

— Steven

FAQ

What is a customer intelligence layer?

• A customer intelligence layer is the synthesis layer that sits above a company's data collection and reporting tools. It connects signals from multiple sources — product analytics, marketing data, sales pipeline, customer support — and surfaces what they mean together. It answers questions no individual tool can answer, because those questions require looking across systems simultaneously.

Why do growth companies repeat bad decisions?

• Growth companies repeat bad decisions because each decision is made from one tool's view of the picture. Without a synthesis layer connecting all data sources, teams see their slice — acquisition, product, support, sales — and make calls based on it. Six months later, a different team sees a different slice and makes a similar call, with no visibility into what the previous decision was or why it failed. The loop continues until the synthesis layer exists to break it.

What is missing from a typical growth stack?

• Most growth stacks have data collection tools (Layer 1) and reporting tools (Layer 2). What's missing is the synthesis layer (Layer 3) — the function that connects what the tools see separately and surfaces the full picture. More tools and more dashboards don't fill this gap. They add to Layers 1 and 2, which already have enough data. The missing layer requires a different kind of function entirely.

How is an intelligence layer different from a dashboard?

• A dashboard reports on one system at a time and answers questions you already knew to ask. An intelligence layer connects multiple systems and surfaces patterns you didn't know to look for. A dashboard shows you what happened. An intelligence layer tells you what it means — across all the systems your stack contains.