I walked into a board meeting prep at a $14M ARR company and asked three VPs what last quarter’s revenue was. I got three different numbers. All three were technically correct.
In This Article
The VP of Sales said $3.8M. He was counting booked deals — signed contracts with a start date in the quarter, regardless of when cash came in. The VP of Finance said $3.5M. She was counting recognized revenue per ASC 606 — what the auditors would bless. The VP of Marketing said $4.1M. He was counting total contract value of deals that closed in the quarter, including multi-year commitments, because that’s what his attribution model tied back to.
The CEO stared at her screen for a long moment and said, “So what’s our actual revenue?” Nobody answered, because the question itself was wrong. There is no single “actual” number until the company decides which definition of revenue is the canonical one for each context. The problem wasn’t that someone was wrong. The problem was that nobody had ever decided who was right.
This is metric chaos. And it’s far more common — and far more expensive — than most founders realize.
Why Metric Chaos Happens
Metric chaos doesn’t come from bad people or bad tools. It comes from success. Specifically, it emerges from the same organic growth that makes a company worth running.
The Organic Growth Tax
When you’re 10 people and $2M ARR, metrics are simple. The CEO knows the numbers because they built the spreadsheet. Everyone uses the same definitions because there’s only one person defining them.
At 50 people and $10M ARR, you have departments. Marketing set up their own analytics. Sales has Salesforce. Finance has NetSuite. Product has Amplitude. Each team configured their tools independently, at different times, with different assumptions. Nobody did anything wrong. They just did it separately.
The result: five tools, five definitions of “customer,” five ways to calculate “revenue,” and zero documentation of why any of them differ.
The Documentation Deficit
In theory, every metric should have a written definition: what it measures, how it’s calculated, what source system it comes from, and who owns it. In practice, I have never once walked into a $5-25M company and found this documentation. Not once in dozens of engagements.
Why? Because documentation feels like overhead when you’re growing fast. And it is — until it isn’t. The tipping point is usually around $8-12M ARR, when the complexity of the business outgrows any one person’s ability to hold all definitions in their head. After that point, every meeting without documented metrics is a meeting where 15 minutes gets burned on “wait, where did you get that number?”
The Tool Configuration Drift
Here’s a specific example I see constantly. A company uses Stripe for billing and has a BI tool (Metabase, Looker, whatever) that pulls from the Stripe API. At some point, an engineer builds a custom revenue model in the app database that calculates MRR differently — it accounts for free trials, or discounts, or usage-based components — because the Stripe number “wasn’t quite right.”
Now there are two revenue numbers. The BI tool shows one. The internal dashboard shows another. Both are sourced from “real data.” Neither is wrong. But they’re different by 8-12%, and every time someone notices the discrepancy, 2-3 hours of someone’s time evaporates investigating why.
Multiply this by every metric in the company — CAC, LTV, churn, NRR, conversion rate, activation rate — and you get a picture of a company where nobody fully trusts any number, but nobody has time to fix any of them.
Departmental Incentives
This one is uncomfortable but real. Different departments have incentives to define metrics in ways that make their numbers look good.
Marketing prefers to count revenue as “influenced” rather than “directly attributed” — it makes their contribution look bigger. Sales prefers to count bookings rather than recognized revenue — it inflates their quarterly number. Finance prefers the most conservative definition — it reduces audit risk. Product prefers active users to paying users — it makes engagement look healthier.
None of these people are lying. They’re each using a defensible definition that happens to align with their incentives. The problem is that nobody has centralized the decision about which definition is canonical for which context.
The Hidden Costs of Metric Chaos
Metric chaos doesn’t show up as a line item on the P&L. It shows up everywhere else.
Decision Paralysis
When leaders don’t trust the numbers, they don’t make data-driven decisions. They make gut decisions — or worse, they delay decisions while someone “gets the right number.” I’ve watched a pricing decision at a $20M company get delayed by six weeks because the team couldn’t agree on the baseline conversion rate. The actual debate wasn’t about pricing strategy. It was about which funnel number was correct.
Estimated cost: Every week of delayed strategic decision is worth roughly 0.5-1% of quarterly revenue in opportunity cost. For a $20M company, that six-week delay cost $150K-$300K in unrealized impact.
Eroded Trust in All Data
This is the most insidious cost. Once people catch the data being “wrong” — meaning different from what they expected — they stop trusting everything. Even numbers that are perfectly accurate get questioned. The data team‘s credibility collapses, and the company reverts to opinion-based decision-making.
I’ve seen this pattern destroy entire data teams. The analysts are producing correct numbers, but because the last report had a discrepancy (due to metric definition differences, not errors), every future report gets interrogated. The analysts spend 60-70% of their time defending their numbers instead of finding insights.
Meeting Time Wasted on “Which Number Is Right?”
Track this in your next leadership meeting. Time how many minutes are spent on: “Where did that number come from?” “That doesn’t match what I see in [other tool].” “Can someone reconcile this before next week?”
In companies with metric chaos, I typically see 15-25 minutes per hour-long meeting spent on meta-discussion about data rather than discussion using data. For a leadership team of 6 people meeting weekly, that’s roughly 1,500 hours per year of combined executive time spent arguing about numbers instead of using them.
At a blended cost of $150/hour for senior leadership time, that’s $225K per year in wasted meeting time alone.
Analyst Burnout From Ad-Hoc “Just Check This” Requests
Every metric discrepancy generates a cascade of ad-hoc requests. “Hey, can you just check the revenue number?” “Can you pull the churn rate from Stripe directly?” “I need the CAC breakdown but using the marketing team’s definition, not finance’s.”
Each “quick” request takes 30-90 minutes. A typical data analyst in a company with metric chaos gets 5-10 of these per week. That’s 10-30 hours per week of reactive work that crowds out the proactive analysis the company actually hired them to do.
This is why your analyst is burned out. It’s not that they’re bad at their job or that you need more headcount. It’s that they’re spending 60-70% of their time as a human reconciliation engine.
The Fix: The 90-Minute Metric Definitions Workshop
The solution is surprisingly simple. Not easy — it requires getting the right people in a room and making decisions — but simple in structure. Here’s exactly how to run it.
Pre-Work (30 minutes, done before the workshop)
Assign one person (analyst, data lead, or fractional CDO) to create a draft inventory of every metric the company uses in leadership meetings, board decks, and departmental reviews. For each metric, list:
- Current name(s) — often the same metric has 2-3 names
- Where it appears (which dashboard, report, or deck)
- Current source(s) — which tool or query produces it
- Known discrepancies — where different sources give different numbers
This inventory doesn’t need to be perfect. It needs to be visible. The act of writing it down reveals 80% of the problem.
Part 1: Align on Core Metrics (30 minutes)
Who’s in the room: CEO, VPs of each department, data lead/analyst.
Start with this question: “If you could only see 10 metrics to run this business, what would they be?”
Write them on a whiteboard. You’ll get 15-20 suggestions. Group duplicates. Debate is encouraged — but the CEO makes final cuts. You should end with 8-12 core metrics. These are the ones that get official definitions.
Typical core metrics for a $5-30M SaaS company:
- Monthly Recurring Revenue (MRR)
- Annual Recurring Revenue (ARR)
- Net Revenue Retention (NRR)
- Gross Churn Rate
- Customer Acquisition Cost (CAC)
- Lifetime Value (LTV)
- CAC Payback Period
- Conversion Rate (trial to paid, or equivalent)
- Active Users (monthly/weekly)
- Gross Margin
Part 2: Define Each Metric (45 minutes)
Go through each core metric and fill in this template:
—
Metric Name: [Official name — one name, no aliases]
Definition: [Plain English description a new employee could understand]
Formula: [Exact calculation with specific terms defined]
Source System: [Where the canonical number comes from]
Known Exclusions: [What is intentionally NOT included and why]
Owner: [One person responsible for accuracy]
Update Frequency: [Real-time / daily / weekly / monthly]
Where It’s Displayed: [Dashboard URL, report name]
Known Limitations: [What this metric does NOT tell you]
—
The “Known Limitations” field is the most important and most overlooked. Every metric has blind spots. MRR doesn’t capture usage-based revenue fluctuations. NRR can mask new customer acquisition problems. Documenting limitations prevents people from over-indexing on a single number.
The most important decision in this section: when two definitions of the same metric exist, the group must choose one as canonical. Not “both are valid” — one is the official number for leadership and board reporting. Other definitions can exist for departmental use, but they must be labeled as such.
Part 3: Assign Ownership and Next Steps (15 minutes)
For each metric:
- Assign a single owner. This person isn’t responsible for computing the metric — they’re responsible for its accuracy. If the number looks wrong, they investigate. If the definition needs to change, they propose the change.
- Set a review cadence. Core metrics should be reviewed for accuracy quarterly. Definitions should be reviewed annually or when the business model changes.
- Decide where the definitions live. Not in someone’s head. Not in a Slack message. In a persistent, searchable, version-controlled document. Notion, Confluence, a wiki, even a Google Doc — anything that lives in one place and is accessible to everyone.
Post-Workshop: The Metric Definition Doc
Within 48 hours of the workshop, the data lead should publish the metric definition doc and share it with the entire company. Not just leadership — everyone.
Then do one more thing: set up automated monitoring. If MRR in Dashboard A differs from MRR in Dashboard B by more than 1%, someone gets pinged. Discrepancies will happen — new tools, schema changes, formula updates. The key is catching them in hours, not months.
What Changes After the Workshop
The first thing you’ll notice: meetings get shorter. When everyone agrees on the definitions, you skip the 15-minute “which number is right?” debate and go straight to “what does this number tell us and what should we do?”
The second thing: analyst productivity jumps. When there’s a documented source of truth, stakeholders stop asking “can you just check this number?” because they can check it themselves. I’ve seen analyst teams reclaim 15-20 hours per week after a metric definitions workshop — that’s like hiring another half an analyst for free.
The third thing: trust rebuilds. Once people see the same numbers in every report, in every meeting, from every tool — they start trusting the data again. And when they trust the data, they use it. Decisions speed up. Arguments become productive (“should we invest in retention or acquisition?”) instead of procedural (“wait, what’s our actual churn rate?”).
The Metric Chaos Maturity Scale
Here’s a quick self-assessment. Where is your company?
Level 1: Tribal Knowledge (most $3-8M companies)
Metrics live in someone’s head. One or two people can explain any number, but nothing is written down. If those people leave, the company loses its data memory.
Level 2: Spreadsheet Era (most $8-15M companies)
Metrics are tracked in spreadsheets and dashboards, but definitions vary by department. Regular disagreements about numbers in meetings. Analyst time is dominated by ad-hoc requests.
Level 3: Documented but Not Enforced (some $15-25M companies)
Metric definitions exist somewhere (a wiki page, a Notion doc) but aren’t consistently used. Some dashboards reflect the official definitions; others don’t. Compliance is voluntary.
Level 4: Governed (rare below $30M, but achievable)
Core metrics have official definitions, single owners, automated monitoring, and regular review cadences. Dashboards are certified. The metric definition doc is a living document that new employees read in their first week.
Level 5: Self-Healing (enterprise-grade)
Metric governance is automated. Schema changes trigger definition reviews. Data contracts between teams prevent drift. The data team focuses on insight, not reconciliation.
Most companies I work with are at Level 1 or 2. The workshop described above gets you to Level 3 in one afternoon. Getting from Level 3 to Level 4 typically takes 2-3 months of sustained effort. But the jump from Level 1 to Level 3 delivers 80% of the value. Don’t let perfect be the enemy of functional.
The Real Problem Isn’t Technical
I want to end with the most important point. Metric chaos feels like a data problem — like something that should be solved by better tools, better dashboards, or better analysts. It isn’t.
Metric chaos is a communication and governance problem. It happens when smart people make independent decisions about how to measure things, without a forum to align those decisions. The fix isn’t a new BI tool. The fix is a 90-minute meeting where adults agree on what words mean.
That’s it. That’s the whole secret. Get in a room, agree on definitions, write them down, assign ownership, and review regularly. The tools, dashboards, and pipelines are downstream of that agreement. Without it, every tool you buy and every analyst you hire is building on quicksand.
Your company probably has six versions of revenue right now. Pick one. Write it down. Make it official. Everything else gets easier after that.
—
*Nick Valiotti is a Fractional CDO who helps $5-30M ARR companies turn metric chaos into metric clarity. The Metric Definitions Workshop is the first thing he runs with every new client — because you can’t build on data you can’t agree on. Learn more at valiotti.com.*
*Want to run this workshop at your company? Download the Metric Definitions Workshop template — includes the facilitation guide, pre-work checklist, and metric definition doc template.*
