Your head of product defines retention as "users active in month N who were also active in month N+1."

Your growth person defines it as "users who paid in month N who paid again in month N+1."

Your finance person defines it as "customers with positive recurring revenue in both periods."

Same word. Three different numbers. Three different realities.

Every meeting becomes a translation exercise. Every comparison becomes useless. You're not actually disagreeing about what happened—you're disagreeing about what you're measuring.

How Definition Drift Starts

It doesn't happen with a bang. It's quiet.

Early stage, you're shipping code. You're not sitting down to formally define metrics. You don't have a data team yet, or you have one person who's drowning. Teams build their own models to answer their own questions. Marketing models monthly active users one way. Product models it another. Finance has a third version in a spreadsheet.

Nobody announces it. The divergence just... happens.

Then someone looks at a dashboard and says, "Huh, that's weird. This chart shows retention up 5% but the other chart shows it down 3%." And you spend an hour in Slack figuring out whose definition is correct. Spoiler: you both are. You're just measuring different things.

Six months in, you have seven definitions of retention floating around your organization. Each team has good reasons for their version. Each version is technically defensible. None of them match.

The Real Cost

Confusion is annoying. But definition drift costs you actual decisions.

You look at a retention chart trending up and decide to double down on onboarding. Three months later, you realize that the version of retention your product team uses doesn't match the version finance uses to track churn. The cohorts don't align. The trend isn't what you thought. You made a decision based on a number that meant something different to someone else.

Or you're in a board meeting. Investor asks, "What's your retention?" You give a number. Next week, your finance person sends a different number. Not wildly different, but different enough to change the narrative. You look disorganized. The investor notices.

Or worse: two teams are optimizing based on the same metric name but different definitions. One team succeeds by their definition, the other team fails by theirs. They're working at cross purposes without knowing it. The company pulls in two directions.

The time cost compounds too. Every report becomes a negotiation. "Is that using the new definition or the old one?" "Can we run that against the product version?" Someone rebuilds dashboards. Someone else questions the numbers. Someone owns a spreadsheet to "correct" the official data.

Why It's Hard to Fix

You can't just declare a definition and move on.

Retention is genuinely ambiguous. Do you count users who logged in once and never returned? Do you count free users the same as paying customers? What if someone churned but came back? Is that retention or resurrection?

Each answer makes sense in context. Product cares about engagement (free users count). Finance cares about revenue (only customers count). Growth cares about cost of acquisition relative to lifecycle value (need both).

So teams don't converge on a definition. They converge on a spreadsheet, a dashboard, and a lot of trust erosion.

The Fix: One Definition, Enforced Everywhere

The fix isn't to have a better meeting or a clearer memo. The fix is infrastructure.

You need a place where you define "retention" once, and every query, every dashboard, every alert uses that definition. Not the Excel version. Not the Looker version that Bob built for himself. The definition.

At Lopus, that's what the semantic layer does. You define retention in plain terms: users active in month N who returned in month N+1. That definition lives in one place. When someone asks "what's our retention?", the system pulls from that layer. Everyone gets the same number, calculated the same way.

The definition can evolve as your business does. You're at Series B and now you want to separate free users from customers in your retention calc? You change the definition once. The next time someone runs the query, they get the new number. No inconsistency. No confusion.

This matters more than you think because it's not really about the number. It's about trust. When everyone knows they're looking at the same thing, you can actually have a conversation about what to do with the number instead of arguing about what the number means.

Why This Matters Now

At your stage, you have two or three data tools and four different definitions of every metric that matters. That divergence looks small. It's not.

It compounds. The bigger you get, the more expensive definition drift becomes. Series B, you have five teams. Series C, you have fifteen. Each team has added their own wrinkle to "retention" or "churn" or "LTV." Now you need someone full-time just enforcing definitions.

Better to build the habit now. Define your metrics. Enforce them. Keep everyone aligned. The teams that do this early end up making faster decisions later because they're not constantly re-translating the same words.

The immediate step: Look at your top five metrics. Ask your team to define each one. You'll probably find at least two different versions of each. That's your signal. You need governance, and you need it before it gets worse.