Marketing Says 10k Signups. Product Says 2k. You Need a Single Source of Truth.

It’s Monday morning. You’re in the weekly business review, walking everyone through last week’s performance. You put up a slide showing 2,000 new user signups, tracked directly from your product analytics tool. You feel good.

Then the head of marketing presents. Her first slide shows a big, bold number: 10,000 new users from the latest campaign, pulled from their ad platform.

The room goes quiet. The CFO looks from her screen to yours, then back again. Someone asks the question you’re all thinking: “So… which number is right?”

Every product manager has lived this moment. It’s a data standoff. You spend the next 30 minutes defending your tracking implementation while marketing defends their attribution model. Nothing gets decided. The meeting ends with a vague action item to “reconcile the numbers.” You just lost an hour of your life to data forensics instead of product strategy.

Your Data Isn't Lying, It's Just Speaking Different Languages

Conflicting metrics rarely come from a single broken query. They come from a dozen tiny, uncoordinated decisions made over several years.

• Different Definitions: Does “user” mean anyone who visits the site, or someone who creates an account? Does “active” mean opening the app, or completing a key action? Marketing and Product often have different answers.
• Different Tooling: Marketing lives in Google Analytics and ad platforms, which use cookies and last-touch attribution. Product lives in Amplitude or Mixpanel, which use user IDs and event-based tracking. They are measuring different things in different ways.
• Different Timestamps: Your product database runs on UTC. The finance team’s report is set to EST. Your weekly report ends on Sunday at midnight, but theirs ends on Friday at 5 PM. These small gaps create massive confusion.

The result is death by a thousand data cuts. You can’t make good decisions because you can’t even agree on what happened.

The Answer Is Not Another Dashboard

The typical reaction is to commission a new, unified dashboard. A "master dashboard" that pulls everything into one place. This is a trap.

You’ll spend two sprints pulling engineers off product work to build complex data pipelines. You’ll hold six meetings to argue about which tool’s definition of a “session” is the correct one. By the time the dashboard is finally built, a new marketing channel will have been added, a key product event will have changed, and the whole thing will be out of date.

You’ve treated a communication problem like a technical one. The goal isn’t to have one perfect chart. The goal is to have one shared understanding.

Create a Metrics Dictionary, Not a Data Pipeline

Stop building. Start writing. Your first step toward a single source of truth is a simple, shared document. Call it a Metrics Dictionary or a Semantic Layer. I just call it "the page we point to when people argue about numbers."

Here’s how to build it.

1. Assign a DRI for Each Core Metric

For every key performance indicator your business tracks, assign a Directly Responsible Individual. This isn’t about blame; it’s about clarity. The PM for the growth team probably owns “New User Signup.” The lead for the engagement team owns “Weekly Active User.” The marketing ops manager owns “Marketing Qualified Lead.” One metric, one owner.

2. Define the What, Where, and How

Create a simple table in Confluence or Notion. For each metric, the owner must define these five things:

• Metric Name: The common name everyone uses. (e.g., Weekly Active User)
• Plain-English Definition: Describe it like you’re talking to a new hire. What does this metric actually represent? (e.g., “A unique user who logged in and performed at least one key action between Monday 12:00 AM UTC and Sunday 11:59 PM UTC.”)
• Source of Truth (Tool): Name the one system where the official number lives. Not two. One. (e.g., “Our internal Periscope dashboard.”)
• Source of Truth (Query/Report): Provide a direct link to the specific query or report that generates the number. No more "I pulled it from the database." Show your work. (e.g., Link to the SQL query or the saved report.)
• Owner: The name of the DRI.

3. Ratify and Publicize

This document is a social contract. The DRI for each metric should get sign-off from their direct stakeholders. The growth PM needs the head of marketing to agree to the definition of a signup. Once ratified, make this page easy to find. Link it in your team’s Slack channel description, your onboarding docs, and every reporting template you use.

From Referee to Records Keeper

The next time you’re in a meeting and someone presents a conflicting number, your job is simple. You don’t have to argue about attribution models or tracking pixels.

You just ask one question: “Is that number from the official source listed in our Metrics Dictionary?”

Suddenly, the conversation changes. It’s no longer your opinion versus theirs. It’s about whether you, as a group, are honoring the agreement you made. This forces the right conversation. If the marketing number is more useful, maybe you need to update the dictionary. If it’s not, the person who brought it is responsible for reconciling it with the source of truth, not you.

Building a metrics dictionary isn’t glamorous. It’s tedious documentation. But it trades hours of weekly debate for a one-time alignment effort. It lets you stop arguing about the numbers and start making decisions with them.