The Hierarchy of Metrics

A board director once put a slide in front of me showing our Rule of 40 had slipped from 44 to 31, tapped it, and asked what we were doing about it this quarter. The honest answer was “nothing about the number itself,” and it took me a few uncomfortable seconds to explain why. Rule of 40 is last year’s growth added to last year’s margin. The things that dragged it down (a lead-response regression back in Q1, discount creep that crept in over Q2) had already happened, were already over, and had already shown up months earlier on operator dashboards nobody in that room had ever opened. Asking the sales floor to “improve Rule of 40” this week is like asking a driver to change the odometer reading by pressing harder on the wheel. The number is real. It is just not a control.

That gap between a metric you can read and a metric you can move is what this chapter is about. The previous chapter (/course/data-governance-quality/) got the bottom layer trustworthy: the data model your whole business measures itself with. Now we stack four more layers on top of it and give each one a job, because the single most common instrumentation mistake I see is a good operator trying to run daily operations on a metric that was built to price the company, not to steer it.

Metrics have an altitude

Revenue Architecture arranges every number a GTM business produces into five layers, and the layers are not a ranking of importance. They are a ranking of altitude: how far a metric sits from the raw event, how many other numbers get compressed into it, and therefore how slowly it moves and how little any one person can do to it directly.

Read the ladder from the bottom up. L1 is the data model itself, the governed Bowtie fields from the last chapter. L2 is the tactical layer: the volume, conversion, and time markers (VM, CR, Δt) the metrics spine (/course/metrics-spine/) hung on each Bowtie stage, plus the cost of each lead and meeting. L3 is where those tactical numbers combine into operator performance: unit economics, a motion’s Growth Formula, rep productivity. L4 is the strategic-financial layer the CFO and CRO live in: ARR, NRR, CAC, FCF margin. L5 is the executive-investor layer: the ratios a board and an acquirer use to value the company.

The rule the ladder encodes: a metric is only as true as the layer beneath it, and only as actionable as it is low. Every L5 ratio is a stack of L1 fields. That is not an abstraction. The reporting field guide I keep for our Salesforce org catalogs 1,390 reporting-relevant fields, and exactly one of them, New_ARR__c, is the canonical gate for net-new pipeline. Get that one field wrong on a few percent of opportunities and your Magic Number is wrong too. You just cannot see the error from the top of the ladder, which is the whole danger of living up there.

The three questions each layer answers

Back in /course/models-systems-loops/ we named the three lenses leadership uses to make an informed decision, the 3 Es. They map cleanly onto the altitudes and tell you what each metric is really asking.

• Effectiveness is the scalable-growth question: are we producing wins fast enough? It lives on the left of the Bowtie, in velocity, volume, and conversion. This is where Magic Number gets its signal.
• Efficiency is the sustainable-growth question: what does a dollar of that growth cost, and is the cost trending down? CAC, CAC Payback, and Burn Multiple are all efficiency compressed into a single ratio.
• Economy is the durable-growth question: is the growth high quality, meaning does it come from retained, expanding, impacted customers? GRR, NRR, and FCF margin answer this one.

The investor metrics at the top are not new information. Each is one of the three Es squeezed into a number a non-operator can act on at the altitude of a board seat. Rule of 40 is the deliberate blend, effectiveness and economy in one line, which is exactly why it is useless as a lever: it deliberately hides which of the two is failing.

Defining the metrics the book only names

Revenue Architecture slots six investor metrics into the top of the hierarchy and, oddly, never gives their formulas or healthy ranges. Operators are expected to know them cold. Here they are, with the benchmark bands that circulate among SaaS investors. Treat the ranges as heuristics, not laws: they shift with segment, stage, and interest rates.

A few of these carry names worth knowing. Rule of 40 is Brad Feld’s 2015 heuristic. LTV:CAC’s 3:1 target is David Skok’s. Burn Multiple is David Sacks’s, and it is the sharpest of the group because it needs no assumptions about lifetime or margin: cash out, new ARR in, done. The Magic Number and CAC Payback are two views of the same efficiency question, one as a ratio and one as a duration, and they will disagree when gross margin is unusual, which is a feature, not a bug.

Operate one so the trailing nature is obvious. Move the two inputs to the Magic Number and watch the ratio cross 0.75. Then notice what you are actually holding: net-new ARR that already closed, and sales-and-marketing spend from a quarter that already ended. Both books are shut. You are reading history at high resolution.

The trap: running the floor on the top of the ladder

The failure mode is not using investor metrics. It is using them where a leading metric belongs. Investor metrics share three properties that make them excellent for valuation and dangerous for operations.

They are ratios, so a healthy top number can hide a sick component. Rule of 40 at 42 looks fine right up until you notice growth collapsed and only a margin cut is holding the line. They are trailing, computed from closed books, so by the time an investor metric moves, the leading metric that caused it moved a quarter or two earlier, when you could still have done something. And several of them are spoofable. Revenue Architecture is blunt about the worst offender: NRR can be inflated by a 5 to 7% list-price increase that masks real logo churn, which is why GRR, harder to game, is the truer read on durability. A rep or a manager staring at a company-level ratio has no move that changes it this week, so the metric generates anxiety instead of action.

The discipline is simple to state. Report up the ladder, steer down it. Give the board L4 and L5 because that is the altitude a valuation is set at. Run the weekly operating rhythm on L1 through L3, because those are the only numbers anyone in the room can actually move before the next meeting.

Handing off

You now have the altitudes and the six investor metrics defined, formulas and bands and all. The three moves that matter: place each metric on its layer before you use it, read every top-line ratio next to the component beneath it, and run daily operations on the leading L1 to L3 numbers while reserving the trailing L4 and L5 ratios for the board and the balance sheet.

Which raises the obvious question this whole module exists to answer: if the metrics you steer with live at the bottom of the ladder, on the Bowtie, then that is what your operating dashboard has to show. Not the valuation ratios that look impressive on a slide, but the volume, conversion, and time tiles a manager can act on before Friday. Building that instrument is next. Continue to Building the Bowtie Dashboard (/course/bowtie-dashboard/).