GTM Foundations - GTM Team KPIs: Definitions, Formulas and What Each One Diagnoses
Every GTM team measures something. Far fewer measure things that change a decision. The test for a useful KPI is simple: if the number moves, does anyone do something differently? If not, it belongs in an archive, not on a dashboard.
This page defines the metrics worth steering by, gives the formula for each, names the diagnostic question it answers, and sets out what has to be instrumented for the number to be trustworthy.
7 min read5 sectionsGTM Foundations
What you'll take away
- Track four layers — efficiency, velocity, quality and retention. A dashboard heavy in one layer will mislead you.
- Any metric that requires manual spreadsheet work will be wrong within a month. Instrumentation is the prerequisite, not the follow-up.
- Segment every metric. A blended win rate across all segments and sources hides the two facts you needed.
- Benchmarks are planning heuristics, not targets. Your own trend line is worth more than any industry median.
The core metrics, with formulas
Twelve metrics cover almost every question a revenue leader asks. Each entry below gives the calculation and what it actually tells you.
- Pipeline coverage
- Open pipeline value in the period ÷ quota for the period. Answers: do we have enough at-bats to hit the number? Most B2B teams plan against 3–4×, but the right multiple is simply the inverse of your historical win rate plus a margin.
- CAC payback period
- Fully loaded sales and marketing cost ÷ (new ARR × gross margin), expressed in months. Answers: how long until a customer pays back what it cost to acquire them? Under 12 months is comfortable; over 24 requires either a pricing change or a cheaper channel.
- Cost per opportunity
- Total demand generation spend ÷ qualified opportunities created. Answers: is our top-of-funnel investment efficient? More actionable than cost per lead, because it survives lead-quality games.
- Stage conversion rate
- Opportunities exiting a stage ÷ opportunities entering it. Answers: where exactly do deals die? The single most diagnostic metric in the set, and useless unless stage definitions are enforced.
- Sales cycle length
- Median days from opportunity creation to closed-won. Use the median, not the mean — one nine-month enterprise deal will distort an average and hide the typical experience.
- Lead response time
- Median minutes from inbound submission to first meaningful contact. Answers: are we reaching people while intent is still live? This is the metric most improved by automation and most degraded by manual routing.
- Win rate by segment and source
- Closed-won ÷ (closed-won + closed-lost), split by segment and acquisition source. Answers: which markets and channels actually work? A blended number is close to meaningless.
- SQL to SQO conversion
- Sales-qualified opportunities ÷ sales-qualified leads. Answers: is marketing's definition of qualified the same as sales'? A low rate here is a definitions problem, not a lead-quality problem.
- Forecast accuracy
- Actual closed revenue ÷ forecast at the start of the period. Answers: can we plan on our own numbers? Below about 85% and every downstream decision — hiring, spend, board guidance — inherits the error.
- Net revenue retention
- (Starting ARR + expansion − contraction − churn) ÷ starting ARR, cohort-based. Answers: does the base grow without new logos? The metric with the largest compounding effect on enterprise value.
- Gross churn
- Churned ARR ÷ starting ARR. Answers: are we losing customers? Track alongside NRR, because strong expansion can mask a serious retention problem in the small-account cohort.
- Time to value
- Median days from contract signature to the customer's first meaningful outcome. Answers: how quickly does the promise become real? Strongly predictive of both churn and expansion, and almost never instrumented.
Reading the four layers together
Individual metrics mislead. The diagnostic power comes from reading them in combination, because most revenue problems have a signature across two or more layers.
| Pattern | Likely cause | First action |
|---|---|---|
| High lead volume, low SQL→SQO | Marketing and sales disagree on "qualified" | Rewrite the qualification definition jointly and encode it as validation rules |
| Good win rate, low pipeline coverage | Top-of-funnel underinvestment, not a sales problem | Shift budget to demand generation before hiring another AE |
| Long cycles, high win rate | You are winning deals that were never competitive — possibly under-priced | Test pricing and qualify harder on urgency |
| Strong NRR, high gross churn | A few large accounts are hiding many small losses | Segment retention by account size and fix onboarding for the small cohort |
| Forecast accuracy below 80% | Stage definitions are subjective | Add exit criteria per stage and enforce them as required fields |
| Fast response time, flat conversion | Speed is fine; routing sends leads to the wrong owner | Audit routing rules and territory assignment |
Metrics to stop reporting
Each of these is easy to measure, easy to improve, and unconnected to revenue. That combination is exactly why they persist in board decks.
- MQL volume
- Trivially inflated by lowering the threshold. Report pipeline contribution instead — it cannot be gamed without generating real opportunities.
- Activity counts
- Calls made and emails sent measure effort, not outcome. They are useful for coaching an individual rep and actively harmful as a team KPI.
- Website traffic
- Relevant only when segmented to ICP-fit visitors and connected to pipeline. Aggregate sessions tell you nothing about whether the right people arrived.
- Blended CAC
- Averaging across channels hides that one channel is excellent and three are wasteful. Always split by channel and segment.
- Number of demos delivered
- A throughput measure that rewards low-qualification demos. Measure demo-to-opportunity conversion instead.
What has to be instrumented for these numbers to hold
A KPI is only as good as the data underneath it. These are the minimum requirements — none of them optional if you want numbers you can defend in a board meeting.
Timestamp every stage transition
Not just the current stage — the full history of when each opportunity entered and left each stage. Without this, cycle length and stage conversion cannot be calculated retroactively at all.
Attribute source at creation and never overwrite it
Store original source as an immutable field alongside a mutable last-touch field. Overwriting original source is the most common reason attribution reporting collapses.
Enforce exit criteria as required fields
A stage that can be advanced without evidence produces a forecast built on optimism. Required fields per stage are unpopular for two weeks and load-bearing forever after.
Log handoffs with an owner and a timestamp
Marketing to sales, sales to onboarding, onboarding to CS. Without a logged handoff you cannot measure SLA compliance, and unmeasured SLAs are not SLAs.
Keep history in a warehouse
CRMs overwrite. A warehouse snapshot lets you answer questions about last year using last year's field definitions, which is the difference between analysis and archaeology.
Automate the calculation
Every metric computed on a schedule from source data, with the definition in version control. If someone has to build the number by hand, it will diverge from the last version within a month.
Cadence: what to look at, and how often
Reviewing everything weekly produces noise and reactive decisions. Match the review frequency to how fast the metric can actually move.
| Cadence | Metrics | Decision it supports |
|---|---|---|
| Weekly | Pipeline coverage, lead response time, stage transitions | Where to focus effort this week |
| Monthly | Cost per opportunity, stage conversion, cycle length | Channel spend and process adjustments |
| Quarterly | CAC payback, win rate by segment, forecast accuracy, NRR | Hiring, pricing, segment focus |
| Annually | Cohort LTV, time to value, full funnel efficiency | Strategy, packaging, market selection |
Frequently asked questions
What are the most important GTM KPIs?
- If you can only track five: pipeline coverage, CAC payback, stage conversion rate, win rate by segment, and net revenue retention. Together they cover whether you have enough pipeline, whether growth is affordable, where deals die, which markets work, and whether revenue sticks.
What is a good pipeline coverage ratio?
- Most B2B teams plan against 3–4× quota. The right number for you is the inverse of your historical win rate plus a margin for slippage — a team converting 33% needs roughly 3×, a team converting 20% needs closer to 5×.
How do you calculate CAC payback?
- Divide fully loaded sales and marketing costs for a period by new ARR won in that period multiplied by gross margin. The result is expressed in months. Under 12 months is comfortable for most B2B SaaS; over 24 usually indicates a pricing or channel problem.
Why is our forecast always wrong?
- Almost always because stage definitions are subjective. If a rep can advance an opportunity without documented evidence, the forecast measures optimism rather than probability. Adding exit criteria as required fields typically improves accuracy more than any forecasting tool.
Should we track MQLs?
- Track them internally as a workflow trigger if useful, but do not report them as a performance KPI. MQL volume is inflated by lowering the qualification threshold. Pipeline contribution measures the same intent without the incentive to game it.
Most KPI problems are instrumentation problems
We build the tracking, warehouse snapshots and automated calculations that make revenue metrics trustworthy — so the board deck and the CRM finally agree.