Finance and the board · Answered by Golden Section from more than 400 B2B software companies observed
Give each department one or two KPIs it can move within weeks, each tied to a line of the P&L or to ARR. Sales owns new ARR against plan and sales efficiency. Marketing owns the share of account-executive pipeline it sources, which we hold at 30% or more. Customer success owns gross revenue retention and time to go-live. Product and engineering own delivery against the committed roadmap and adoption of what shipped. Finance owns the close date and the accuracy of the cash forecast. Every KPI needs one owner, an acceptable range and a written action for when it leaves the range. To tie KPIs to the P&L, book costs where people actually spend their time, so the sales, product and cost of revenue lines tell the truth. Start by writing one sentence per KPI naming the P&L line it moves.
A KPI belongs to a department only if that department can move it, someone is accountable for calculating it, and a reading outside its range triggers an action already written down. Two per leader; five to fifteen on the company's weekly scorecard.
| Metric | Value | What it means | Source |
|---|---|---|---|
| KPIs per leader | no more than 2 | KPIs each accountable leader owns, each with a range and an if/then actionchosen in a meeting where each leader defends why the KPI supports the company's crucial number | Golden Section playbookKPI Dashboard Creation |
| Weekly scorecard | 5 to 15 numbers | weekly, leading measures, each with one owner and a targetif populating it takes one person more than an hour on Monday, it is the wrong scorecard | Golden Section playbookExecution Operating System |
| Marketing-sourced AE pipeline | 30%+ | share of account executive pipeline sourced by marketingthe KPI we give marketing, alongside AEs' own relationship building and outbound | Golden Section operating viewGolden Section operating view |
| Sales efficiency target | 0.70 or less | prior-period S&M expense ÷ new ARR booked, including upsellsthe sales leader's ratio; it only reads true if sales and product costs are coded correctly | Golden Section playbookSales Efficiency Ratio |
| Gross revenue retention | 95%+ | starting ARR retained after churn and contraction, excluding expansionthe level needed to sustain top-tier cash-flow margins; the customer success KPI with the most value attached | Golden Section, publishedInvesting in Software addendum |
| Contribution margin, GS definition | revenue − cost of revenue − CS and existing-account S&M − G&A | excludes new-account sales and marketing and new-product R&Dshows what the existing book earns before growth investment; shared people allocated by time spent | Golden Section operating viewGolden Section operating view |
A KPI is useful when it is near real time, has an obvious action attached, is cheap to compute and has a named owner. Most department dashboards fail the second test. They report activity with no if/then logic, so a bad week produces a discussion rather than a response. The KPI dashboard play caps each leader at two, and the execution operating system puts the weekly subset on one scorecard read in one meeting.
The link to the P&L is cost coding. Each department's KPI should move a line its own costs sit on, which only works if the lines are honest. Engineers who spend half their week on implementation belong partly in cost of revenue; product salaries booked to sales and marketing flatter gross margin and wreck sales efficiency. A poor burn multiple sometimes turns out to be nothing more than operating expense in the wrong line. Book hosting, support and implementation labor to cost of revenue, as the P&L and buyers expect. Allocate shared people by time actually spent rather than spreading them by formula. Contribution margin then becomes simple: revenue less cost of revenue, less customer success and the sales and marketing that serves existing accounts, less G&A. Leave out new-account sales and marketing and new-product R&D, which is most of both.
The figures are fictional. A company at $8M in annual revenue reports 82% gross margin and a sales efficiency ratio of 1.1, and its product team insists the R&D budget is too small. Recoding by where people spent their time moves six implementation engineers out of R&D into cost of revenue and two solutions engineers out of R&D into sales. Gross margin falls to 74%, sales efficiency rises to about 1.2, and R&D turns out to be smaller than the board thought. Each change gives a leader a KPI they can now own honestly: customer success takes time to go-live, sales takes efficiency on the restated basis, and product takes adoption of the last three releases.
A team without separate functional leaders should run five company numbers rather than departmental KPIs. And measures that cannot move weekly, such as retention, belong on the monthly review rather than the weekly scorecard.
From the Golden Section mistakes list, each paired with the play that prevents it.
A KPI without one accountable owner is a number everyone watches and nobody moves.
Departments drift toward measuring what is easy, such as activity and cost per lead, instead of the output the P&L pays for.
KPIs that trigger no decision create reporting work that changes nothing.
In the order we would run them. Each is on its own page, most with a free Excel template.
Chooses no more than two KPIs per leader, each with a range and a written action.
Puts the weekly subset on one scorecard with an owner on every line and reads it in a fixed meeting.
Places revenue and costs in the right lines so each KPI connects to a number that means something.
Takes the analysis down to the customer, where contribution by segment becomes visible.
Gives every department a budget owner and ties incentives to the same numbers.
Executive plays The dashboard, meeting and budget plays that turn departmental numbers into decisions.
Map each KPI to the line it moves: new ARR and sales efficiency to revenue and sales and marketing, gross retention to revenue, time to go-live to cost of revenue. Then make sure costs are coded where people actually work, because a KPI tied to a mis-stated line rewards the wrong behavior.
Our definition: revenue less cost of revenue, less customer success and the sales and marketing that serves existing accounts, less G&A. Leave out new-account sales and marketing and new-product R&D, which is most of both, because those are investments in growth rather than costs of the business you already have. The result shows what the existing book earns and which product lines or segments fund the company and which consume it.
Allocate shared people by the time they actually spend, using a short quarterly time survey if needed. Formula allocations by revenue or headcount make every line look the same and hide the decisions the numbers were supposed to inform.
Reviewed by Dougal Cameron, CEO & Co-Founder on 2026-10-22. Golden Section observations are labeled separately from external benchmarks and illustrative arithmetic.