Beyond Rule of 40: The SaaS Metrics Boards Actually Read in 2026

Ten years ago the Rule of 40 was the number every SaaS board reached for first: growth rate plus operating margin should sum to at least forty. It survived the growth-at-all-costs era because it made room for high growth with heavy losses. It survived the interest-rate reset because it made room for slowed growth with restored margins. It did not survive 2025.

Two changes broke it. First, cost of goods sold — historically a small line item for a SaaS business — became large and volatile for any company with a serious AI feature. Second, the market matured to the point where "above 40" became a table stake, not a differentiator. Boards need finer instrumentation.

This is the framework replacing it. For the shorter treatment of why Rule of 40 broke, see our Rule of 40 is Dead audit. This piece is about what to build in its place — the specific report structure a serious CFO or CEO should be running for their board in 2026.

The three-part framework

The composite picture that serious SaaS boards now use has three moving parts. They complement each other; no one of them alone tells the story.

1. Growth-Efficiency triangle — plots growth rate against net dollar retention against burn multiple. Named and popularised in slightly different forms by Bessemer, OpenView, and Meritech. The triangle is a mental model rather than a formula: a company growing thirty percent, retaining net revenue at one hundred and twenty percent, and burning less than one dollar of new capital per dollar of net-new ARR is in a fundamentally different place than a company hitting the same growth rate through paid acquisition with flat retention. Boards look at the shape of the triangle over time. Shifts matter more than absolute values.

2. AI-COGS disaggregation — the single largest analytical change in board packages over the last two years. Serious boards now see cost of goods sold split into hosting, third-party APIs (which is where the AI spend lives), support, and other. They see cohort gross margins — early-cohort customers versus late — because AI-heavy usage patterns are often concentrated in the newest cohorts. They see model-mix reporting because different model routes (frontier versus small, direct versus routed) have different unit economics.

3. Rule of X — popularised by Altimeter and others, this is a family of formulas that weights growth more heavily at high growth rates and less heavily at low growth rates. The common shape: a dollar of growth is worth roughly two dollars of margin at the top of the growth curve and roughly equal to a dollar of margin as growth matures. Rule of X is a better fit for the current environment because it reflects the observed valuation multiples: high-growth companies still command a premium; slow-growth companies are being priced on profitability. The Rule of 40 assumed those were substitutable. The market says they are not.

Boards do not pick one framework and use it in isolation. They use all three at different altitudes.

The disaggregated P&L

The most consequential change in the actual reporting artifact is the P&L itself. The typical pre-2023 SaaS P&L had cost of goods sold as a small consolidated line — perhaps 20% of revenue, growing predictably. That line is now the most-scrutinised in the deck for any AI-touching company.

The disaggregation a serious board expects:

Hosting infrastructure. AWS, GCP, Azure, or specialised inference providers. The compute cost that is not AI-model-specific — API servers, databases, storage, CDN. Growing with usage but relatively predictable per-user.

Third-party AI APIs. OpenAI, Anthropic, Google, or other model provider spend. The most volatile line in the P&L in 2026 — it can move ten percent quarter-over-quarter based on model-provider pricing changes, usage-pattern shifts (users engaging more with agentic features), or a new feature launch that lands more inference on a heavier model tier.

Support cost. Human customer support, developer relations, and the fraction of engineering that is customer-facing. Grows with the customer base but at a decreasing rate.

Other COGS. Payment processing, third-party data feeds, contractor costs directly tied to serving customers.

A modern board package shows this disaggregation month-over-month, with a per-cohort view: what is the gross margin of customers who joined last quarter versus customers who joined two years ago? For AI-heavy products the newest cohorts often have materially different (usually worse) gross margins because they use the newest, most inference-heavy features harder.

Cohort economics with AI COGS

The cohort view surfaces problems the aggregate view hides. Three specific patterns worth watching:

AI-heavy cohorts landing at lower gross margin. New users engage disproportionately with the newest features, which are usually the most AI-heavy. If gross margin on the last two cohorts is 20 points below gross margin on customers older than two years, the business is bleeding margin on new revenue even if the aggregate looks stable.

Model-mix drift. As a product matures, the mix of model routes it uses drifts. A product that started with 90% frontier-model calls may drift toward 60% frontier / 40% smaller model over time as the team optimises. That drift shows up in gross margin trajectory only if you disaggregate.

Provider-pricing exposure. If a single model provider represents more than 40% of a company's inference spend, the board should know. A pricing change from that provider is a first-order risk to gross margin.

None of this is complicated to report. It is expensive to instrument the first time and cheap to maintain after that. Companies that build it before their Series B avoid the painful retrofit later.

The report structure that boards actually look at

Composite: at the beginning of a board meeting, the CFO shows a single page with growth rate, net dollar retention, gross margin (disaggregated as above), and rule-of-40 plus rule-of-X. Everyone glances at all of them. The rest of the meeting is spent on the specific decisions — pricing changes, capital allocation, hiring plan — that flow from what the numbers say.

The single-number simplicity of the Rule of 40 is gone. In exchange, boards have a finer-grained view of what is actually happening in the business. Every CFO I have spoken to who has moved to this framework has said the same thing: they do not want to go back.

What to build if you are a Series A or B founder

Two priorities, in order:

One: instrument the AI-COGS disaggregation now. Build the reporting infrastructure to split hosting, third-party AI APIs, support, and other. Automate it so it shows up in every monthly P&L without manual work. This is the single most valuable investment you can make in your finance stack this year.

Two: build cohort-level gross margin reporting. Same disaggregation, but grouped by customer signup quarter. You will discover things about your business that the aggregate view hides.

Everything else in the modern SaaS framework flows from these two. You cannot do the growth-efficiency triangle or Rule of X analysis well without a clean disaggregated P&L feeding it.

Sources

• Bessemer State of the Cloud reports — bvp.com/atlas
• Bessemer Rule of 40 primer — bvp.com/atlas/the-rule-of-40
• OpenView SaaS benchmarks — openviewpartners.com
• Altimeter Rule of X analysis — altimeter.com