Data teams love their own metrics. Match rates, fill rates, pipeline uptime. And leadership nods politely. Because none of those words sound like money.
These three terms are the translation layer. They’re how data work gets measured in the units the business actually runs on. Let’s define them properly 👇
30-Second Summary
Business metrics turn customer data into revenue arithmetic: what a customer is worth over time, how fast you’re losing them, and how behavior differs by the group they joined in.
What This Category Covers
- Worth: customer lifetime value
- Loss: churn rate
- Behavior over time: cohort analysis
📌 Quick take: All three metrics share one dependency nobody mentions: clean, complete customer records. Garbage data makes every one of them fiction.
The Three Metrics, Walked Through
Customer lifetime value prices the relationship. The page takes you from napkin formula to discounted, segmented models. Then it adds the CLV:CAC and payback lenses that turn the number into acquisition and retention decisions. Its operating principle sets the tone for the whole folder: a used rough number beats an admired precise one.
Churn rate measures the leak. You get customer vs revenue churn, and why they diverge. You get the denominator traps that corrupt comparisons, plus gross vs net revenue churn with negative churn as the compounding grail.
Then come the diagnostic cuts: by cohort, segment, tenure, and reason. They turn a smoke alarm into a repair plan. And the prediction section covers the shift from autopsy to early warning.
Cohort analysis supplies the method both neighbors depend on. Group customers by vintage. Then read the retention grid like a practitioner: down the columns for vintage comparison, along the diagonal for calendar events, with healthy suspicion of young cohorts. Revenue cohorts draw the empirical CLV curve. And behavioral and channel cohorts answer which KIND of customer works.
How the Three Interlock
They form one system. Cohorts reveal how retention actually behaves. Churn summarizes the leak those curves describe. And lifetime value prices what the leak costs, and what fixing it is worth.
Watch a change to onboarding travel. It shows up first in the newest cohort’s early months, then in the churn trend, then in CLV. Same event, three instruments, three time horizons. Teams that read them together stop being surprised by their own business.
And all three share one unglamorous dependency: clean, unified customer records. Duplicate accounts smear cohorts, misstate churn denominators, and split lifetimes across false identities. That’s why the deduplication and enrichment disciplines are silently load-bearing under every number this folder defines. Fix the records and the metrics start telling the truth. The reverse order doesn’t exist.
Using This Folder
Formalizing your metrics? Here’s the practical sequence. Define churn’s conventions first: denominators, statuses, reactivations. One page, frozen. Build the cohort grid second, since it needs only start dates and activity. Then compute segmented CLV third, because it needs the other two.
One quarter of part-time effort. And the business gains a shared numerical language that survives leadership meetings. Which is, in the end, what business metrics are for.
Questions This Category Answers
“How much can we afford to spend acquiring a customer?” The CLV page’s CLV:CAC and payback lenses turn that from debate into arithmetic. Segmented, because averages mislead exactly where the money is.
“Is our retention getting better or worse?” You can’t answer that from a blended average. The cohort grid separates vintages, so improvement or decay is visible per generation of customers. Calendar events show up as diagonal stripes.
“Which customers are we losing, and why?” The churn page’s diagnostic cuts (by segment, tenure, reason, and ICP fit) plus the involuntary-churn split turn one alarming number into a ranked repair list.
“Are the newer customers better?” Compare cohorts down the column. At equal relationship age. It’s the only honest way to grade acquisition changes, onboarding revamps, and pricing moves.
“Why don’t our numbers match finance’s?” Usually definitions and identity: unstable metric conventions and duplicate customer records. The folder’s shared prerequisite (frozen definitions on clean, deduplicated records) is where reconciliation actually happens.
Three metrics, one discipline: revenue relationships measured honestly over time. Teams fluent in this folder stop arguing about whether things are working. The grid, the rate, and the value answer together.
How This Category Connects to the Rest of the Wiki
These three metrics sit at the end of a long supply chain the rest of the wiki describes. The customer records they compute over live in the CRM and reconcile against ERP revenue truth. Their identity hygiene comes from deduplication and matching. Their completeness comes from enrichment. And their freshness comes from the sync and pipeline disciplines.
The analytical machinery is borrowed too. Cohort grids are analytics method applied to retention. Churn prediction is machine learning applied to behavioral signals. And every one of these numbers reaches decision-makers through BI surfaces, where definitional governance decides whether meetings argue or decide.
Which yields this folder’s quiet lesson. When leadership asks why the metrics can’t be trusted, the answer is almost never in this folder. It’s upstream, in identity, completeness, freshness, or definitions. So the fastest path to trustworthy business metrics is unglamorous data work that never appears on the dashboard it rescues.
Start Here If You’re New
New to revenue metrics? Read churn first. It’s the most intuitive, and its definitional traps teach the discipline all three pages share. Cohorts come second, because the grid is the tool you’ll actually open weekly. CLV comes last, once the other two give it honest inputs.
A note on scope. This folder deliberately holds the three metrics with the widest reach. Adjacent numbers (CAC, NRR, payback, activation rates) appear inside these pages where they naturally connect, rather than as separate entries. If the folder grows, it will grow the way the metrics earn trust: one well-defined, well-evidenced term at a time.
The standing invitation across all three pages: compute the rough version this week. Every sophisticated metrics practice began with a napkin number someone was brave enough to publish.
And a final calibration note. Benchmarks for these metrics vary so widely by market, motion, and price point that external comparisons mislead more than they inform. Your own trend is the benchmark that matters, quarter over quarter, cohort over cohort. Compute consistently. Compare internally. And let the folder’s definitions keep the comparisons honest.
Frequently Asked Questions
What are business metrics?
Numbers that measure a business in revenue terms: what customers are worth, how fast they leave, and how behavior changes by vintage. They translate data work into the units leadership actually runs on.
How are CLV, churn, and cohort analysis related?
Cohorts show how retention actually behaves, churn summarizes the leak, and CLV prices what the leak costs. One event (say, an onboarding change) shows up in all three, on three time horizons.
Why do business metrics disagree with finance’s numbers?
Usually unstable definitions and duplicate customer records, not calculation errors. Freeze the conventions and deduplicate the records, and reconciliation mostly takes care of itself.
Which metric should a team compute first?
Churn first, cohorts second, CLV last. Each one feeds the next. A rough version of all three this quarter beats a precise version of one next year.