Measurement

45 metrics. The formula, and what goes wrong.

Each one gives the calculation in the first screen, performs it on real numbers, and names the definition mistakes that make the number lie. Written against what people actually search, not against the topic.
Book a discovery call Get the Weekly Research
01 / Why these read differently

Most metric pages define. These ones decide.

A definition is the cheap part, and it is where most glossary content stops. The expensive part is the cohort rule, the exclusion, the denominator nobody agreed on. That is where reported numbers go wrong, so that is what these pages argue.

Every page carries a worked example with real figures. If a page cannot perform its own calculation, it has not explained it.

02 / What the book is worth

Size and price of the recurring business.

ARPA ARPA divides recurring revenue for a period by the active accounts in that period. It is an average per company, while ARPU averages per individual user, and in B2B the two diverge sharply. A blended ARPA across segments usually describes no real customer, so read it by segment before acting on it. ARPU ARPU divides recurring revenue for a period by the active users in that period. ARPA is the same calculation at the account level, and in B2B SaaS the two diverge sharply. ARPU moves with pricing, packaging, and mix, so the trend needs a driver breakdown before it means anything. ARR ARR is the annualized value of active, committed subscriptions, excluding one-time fees, services, and usage overage. It is not GAAP revenue: ARR is forward-looking and normalized, revenue is what was actually recognized in a period. CARR adds contracted-but-not-yet-live accounts. Report one and label it. Billings Billings describes amounts invoiced during a period; calculated billings is commonly derived from revenue plus the change in deferred revenue, sometimes adjusted for contract assets. It can illuminate subscription sales and cash timing, but contract terms, seasonality, and large renewals make it unsafe as a standalone growth measure. Bookings Bookings is the value of customer contracts signed in a period, recorded at signature under a documented policy. It measures selling output before any invoice or delivery. Billings tracks invoiced amounts and revenue tracks delivered value, so the three diverge by design and must never be swapped. Total Contract Value (TCV) Total contract value is the full committed value of a customer contract across its term: recurring fees times the term length plus one-time fees. ACV annualizes the recurring portion, and ARR describes the run rate. Mixing the three inflates growth stories and breaks deal-size comparisons.
03 / Whether it is holding

Retention, expansion, and what a customer is worth over time.

Churn Rate Churn is customers or revenue lost in a period divided by what you started with. Logo churn and revenue churn answer different questions and diverge sharply when small accounts leave. The denominator is the decision that matters: start-of-period is standard, and mid-period additions must be excluded or growth will flatter the number. CLTV to CAC CLTV to CAC compares lifetime value to acquisition cost, and 3:1 is the convention rather than a law. Its blind spot is time: the ratio treats value arriving in year five identically to value arriving in month three, so a healthy ratio can sit on top of a cash problem. Always read it with payback period. Customer Lifetime Margin CLM subtracts cost of revenue from lifetime revenue and stops there. It sits between lifetime revenue and lifetime profit, and it is useful for comparing segments with different delivery costs. It is not a profitability measure, because acquisition and retention spend are still outside it. Customer Lifetime Profit Customer lifetime profit estimates the discounted contribution a customer or cohort produces over the relationship. Start with revenue, subtract variable product, service, support, and retention costs, model survival by period, and discount future cash flow. Customer Lifetime Revenue CLTR is revenue per customer per period divided by churn, with no margin applied. It is a legitimate top-line planning figure and a bad input to any economics decision, because it ignores the cost of serving. Substituting it into a CLTV to CAC ratio inflates the result by the whole cost of revenue. Customer Lifetime Value CLTV is margin per customer per period divided by churn rate. Three assumptions decide it: which churn, whether margin or revenue, and whether future value is discounted. Using revenue instead of margin and an optimistic churn rate can triple the result, which is why most published CLTV is not usable. Cap the horizon. Customer Retention Cost CRC is the fully loaded cost of retention divided by the customers retained. The definition question is which share of customer success, support, and account management counts as retention rather than expansion, and the split is a judgement you must state. A rising CRC with flat retention usually means you are buying retention that targeting should have provided. Customer Retention Rate Retention rate is customers kept divided by customers at the start, excluding anyone acquired during the period. It is the inverse of logo churn and says nothing about revenue: a business can retain 95% of customers and lose 20% of revenue if the leavers were large. Always report it beside a revenue-based measure. Customer Satisfaction Index (CSI) A customer satisfaction index combines satisfaction scores across weighted relationship attributes into one 0 to 100 number. It differs from CSAT, which rates a single interaction, and from ACSI, the national cross-industry benchmark. The index is comparable only when attributes, weights, and survey wording stay fixed. Lifetime Value to Lifetime Margin CLTV and CLM both apply gross margin to lifetime revenue, so in most constructions this ratio is 1.0 and carries no information. A ratio other than 1.0 means the two numbers use different churn, different horizons, or different cost definitions, which is worth finding before either appears in a decision. Lifetime Value to Lifetime Profit Lifetime value applies gross margin; lifetime profit also subtracts the cost of retaining and serving that customer over their life. The ratio between them isolates the cost of keeping customers, so a widening gap means retention is getting more expensive per customer even when churn looks stable. Lifetime Value to Lifetime Revenue Lifetime value divided by lifetime revenue is gross margin, expressed over a customer's life. It carries no new information, which is exactly what makes it useful: if it does not match your reported gross margin, one of the two numbers is built on different assumptions and needs reconciling. Lifetime Value to Retention Cost Lifetime value divided by retention cost is the retention analogue of the CLTV to CAC ratio: how much value each dollar of retention spend protects. A falling ratio with stable churn means retention is getting more expensive to hold, which is usually a targeting problem surfacing before churn does. MRR Churn MRR churn is recurring revenue lost in a month divided by starting MRR, and it must include downgrades. Excluding contraction is the standard error and it hides a customer base that is shrinking without anybody leaving. Gross MRR churn measures leakage; net includes expansion and is a different number. Net Dollar Retention NDR is what a cohort of existing customers is worth now against what they were worth at the start, including expansion, contraction, and churn, and excluding every new customer. Above 100% means the base grows without acquisition. The number is only comparable when the cohort window and the treatment of downgrades match. Net Dollar Retention Net dollar retention, also called net revenue retention, compares ending recurring revenue from an opening customer cohort with that cohort's starting revenue after expansion, contraction, and churn. New-customer revenue stays out of the calculation.
04 / How well selling converts

Rates and speed through the funnel.

Close Rate Close rate divides deals won by the qualified opportunities or leads worked in a cohort, including the ones that never reached a decision. Win rate divides won by won plus lost only. The two answer different questions, and swapping denominators quietly moves the number by half. Conversion Rate Conversion rate is any step-to-step ratio, so a number without both steps named is not a measurement. Track the chain rather than a headline figure, keep cohorts closed, and read the rates in sequence to find the constraint. A blended funnel rate hides the one step that is actually broken. Pipeline Velocity Pipeline velocity estimates expected revenue flow by multiplying qualified opportunities, average deal value, and win rate, then dividing by sales-cycle length. It is useful only when the cohort and stage definitions stay consistent. Time to Close Time to close measures elapsed days from a documented starting stage to closed-won. The mean is distorted by a few long deals, so report the median alongside it, and never compare series that start the clock at different stages. It is the denominator of pipeline velocity. Win Rate Win rate is closed-won opportunities divided by all closed opportunities in a cohort. The number is meaningless without a fixed stage-entry definition, a closed cohort rather than open pipeline, and clarity on count versus dollar basis. Diagnose it by segment, stage, and loss reason.
05 / Whether the economics work

Cash, margin, and payback.

Free Cash Flow Free cash flow is cash from operations minus capital expenditures, and FCF margin divides it by revenue. In SaaS the number is flattered by upfront annual billing, so read it with deferred revenue movement. FCF margin is one of the two common profitability inputs to the Rule of 40. Gross Margin Gross margin is revenue minus cost of revenue, over revenue. The formula is trivial and the definition is not: hosting, support, customer success, and professional services delivery all belong in cost of revenue, and leaving any of them out inflates the margin. State the line items or the number cannot be compared to anything. Operating Margin Operating margin is operating income over revenue, after cost of revenue and all operating expenses. In SaaS it is suppressed by growth: acquisition spend hits this period while the revenue it buys arrives over years. A negative operating margin alongside strong unit economics is a spending choice, not a broken business, and the way to tell them apart is payback period. Payback Period Payback is acquisition cost divided by monthly gross margin per customer, expressed in months. It is the cash-flow metric the CLTV ratio cannot give you, because the ratio has no time dimension. Use margin rather than revenue, and read it by segment: a blended payback averages motions that finance completely differently. Sales Efficiency Sales efficiency divides new recurring revenue by the sales and marketing spend that produced it. The result depends entirely on two choices: which spend counts, and how much lag you allow between spend and revenue. State both or the number is not comparable to anyone, including your own prior quarters.
06 / Also measured
Cross-Sell Rate Cross-sell rate is the share of customers who bought a different product in a period. It is not upsell, which is more of the same product. Keeping them separate matters because they are driven by different work: cross-sell by product fit and awareness, upsell by usage growth. Bundling breaks the metric entirely, because a bundle sale is neither. CSAT CSAT is the share of respondents who rated an interaction positively. It measures one transaction, usually a support ticket, and does not measure the relationship. Teams that use it as a health signal get a number driven by ticket volume and timing. Ask immediately, report the response rate, and keep the scale fixed. Customer Acquisition Rate Acquisition rate is new customers in a period divided by customers at the start. It measures speed, not cost and not quality. On its own it rewards volume, so read it beside acquisition cost, first-year retention, and average revenue per account, any of which can make a rising acquisition rate a bad sign. Customer Effort Score CES asks how much effort a customer had to expend to get something done. It predicts repeat behaviour better than satisfaction does, because reducing friction retains more reliably than delighting. Watch the scale direction: on some versions a high score is good and on others it is bad, and teams invert it constantly. Customer Health Score A health score is a model predicting renewal, so it is only worth having if it has been back-tested against customers who actually churned. Most are weighted by intuition and never validated, which produces a green dashboard in front of a churn quarter. Build it from behaviour rather than sentiment, and check its accuracy every quarter. Customer Success Metrics Hold customer success to four numbers: net dollar retention, gross revenue churn, retention cost per customer, and a health score that has been back-tested. Everything else is activity. The common failure is a dashboard of touchpoints and QBRs completed, which measures effort and predicts nothing. EBITDA EBITDA is operating income with depreciation and amortisation added back. In software that add-back is contentious, because capitalised development amortisation is a real cost of the product. Adjusted EBITDA goes further and adds back stock compensation and one-offs, at which point it is not comparable between companies without the reconciliation. Expansion Rate Expansion rate is additional recurring revenue from existing customers divided by their starting revenue, excluding churn and contraction. Keeping it gross is the point: netted against churn it becomes net dollar retention and stops telling you whether the expansion motion works. Report both. NPS NPS is the percentage of promoters minus the percentage of detractors, on a 0 to 10 scale. Because passives are discarded, very different distributions produce identical scores. The verbatim follow-up is worth more than the number, and who you sampled matters more than either. Referral Rate Referral rate is customers who produced at least one referred opportunity divided by eligible customers. The hard part is attribution: link tracking undercounts badly because most referrals happen in conversation. Ask on the form and accept self-reported source as the primary signal, with tracked links as a floor rather than the truth. Revenue Run Rate Run rate multiplies one period's revenue out to a year. It assumes nothing changes, which is never true, so it is a snapshot rather than a forecast. Annualising a month that included one-time revenue is the classic error, and it is why run rate and ARR diverge. Upsell Rate Upsell rate is the share of eligible customers who bought more of the product they already have: seats, tier, or volume. Exclude automatic usage-based increases, because those are pricing working rather than a sales motion. Customers already on the top tier are not eligible, and leaving them in the denominator punishes success. Win-Back Rate Win-back rate is churned customers who returned divided by churned customers eligible to return, inside a defined window. Without the window it drifts upward forever, because the pool of past customers only grows. Segment by churn reason: price and timing churn wins back, product-fit churn almost never does.
07 / Where these sit

Measurement is one of six pillars.

A metric only means something inside an operating model. These pages sit under the Measurement pillar of the Tenbound Pipeline Architecture Standard, beside Market, Signal, Message, Motion, and Mastery.

More of the library is being rebuilt. The playbooks cover the procedures these numbers measure.

Who we built this with

A decade of CIENCE sales development engagements.

Next step

Book a discovery call.

Thirty minutes with an account executive who runs the program. We ask how pipeline gets made at your company, and give you our first read on where the problem sits. Nothing is quoted or signed.

Book a discovery call Score yourself first