A founder walked me through their board deck once: thirty metrics, an hour of nodding, and an ending where nobody decided anything about money. The deck had been built to survive questions, and it did exactly that.
I report three numbers to every founder I work with. Three is the count that survived, because a fourth number never once changed what anybody did the following Monday.
How the deck got to thirty
Reporting bloat has an honest origin. Somebody challenges a marketing figure in a meeting, so the marketer adds a slide that would have answered it. Somebody asks about lead quality, another slide. Over a year the deck accumulates every chart that ever ended an argument.
Then each function contributes the number that flatters it. Paid media brings cost per click, which fell. Lifecycle brings open rates, which rose. Sales brings pipeline created, which is up and mostly imaginary.
Every chart is true. None of them are comparable. The room reads thirty true things and has no way to rank them, so the hour ends with a founder who feels marketing is busy and a marketer who leaves without a budget.
A board deck with thirty numbers is a defense document. Three numbers force a decision.
The bill lands on whoever runs marketing. Standing in a room able to produce evidence for any question is a weak position, because the volume of the evidence is what makes it unreadable.
The strong position is being the person who says which number should decide this, and then living with the decision when it goes against them.
The filter: does it force a decision
The test I apply is easy to state and unpleasant to apply. A metric earns a permanent place only if a change in it forces a specific action. If the number could move 20% in either direction and the company would do the same thing next week, it is context, and context belongs in an appendix.
Run that filter across a mature reporting stack and three survive. They survive because they answer three questions that cannot be swapped for one another: whether growth can be bought profitably, how fast the money comes back, and which part of the machine produced this month's growth.
One: LTGP:CAC, and whether to spend more
The first number sets the gross profit a customer generates over their life against everything it costs to win them, marketing spend and selling effort included. I use lifetime gross profit rather than revenue because revenue rewards whichever line moves the most volume, whatever it keeps.
The full derivation, including why the denominator has two halves, is its own article. At Leverage Companies I ran the ratio across three business units with completely unrelated economics, a lender, an education business, and a wholesaling operation, and it was the only figure that put them on one ranked list a CEO could act on.
What it forces is the spend-more decision. Comfortably above the line, the company is buying gross profit at a discount and should buy more of it. Below the line, additional spend digs the hole faster, and whatever needs fixing sits above the ad account.
If you want your own figure before the next board meeting, the LTGP:CAC calculator here will produce it in a couple of minutes from numbers you already have.
One warning about this metric in a board setting. A healthy blended ratio hides a great deal, so compute it per acquisition source and per product line before you report it. Companies get talked into scaling an average that one strong channel was quietly holding up.
Two: payback period, and how hard to press
The ratio says whether the trade is a good one. It says nothing about when the money comes home, and a company can run out of cash making profitable trades if they settle slowly enough.
CAC payback period is the number of months it takes the gross profit from a cohort of customers to cover what you paid to acquire them. Two companies can post identical efficiency and live in different worlds: one recovers its cost in four months and spends the same dollar three times a year, the other waits eleven months and needs a balance sheet to bridge the gap.
Payback period is the permission slip. It tells a founder how hard they can press spend before the bank balance votes.
Benchmarks help a founder calibrate before they panic. Bessemer's cloud guidance is to target payback under 12 months selling to small business, under 18 for mid-market, and under 24 for enterprise.
The operational use of this number is aggression. When payback lands inside a quarter, I will push spend hard, because January's money is back before the annual plan has finished being argued about.
When payback runs past three quarters, every incremental dollar is a financing decision, and the founder should be talking to a bank before talking to a media buyer. Same ratio, opposite instruction.
It also settles the argument about annual plans, which is why the pricing lever outperformed every campaign change around it at FX Replay. Collecting a year of subscription cash up front compresses payback toward zero and turns the acquisition budget into a loop that refills itself.
Three: new revenue by source, and where it goes
The first two numbers judge the machine as a whole. Neither tells a founder which part of it is doing the work, and that is the question that decides where next quarter's budget lands.
So the third is this month's new revenue, broken into the parts that produced it. Paid acquisition. Lifecycle and email. Partners and affiliates. Pricing and plan mix. Four buckets that add to the total, with the prior month sitting beside them.
Revenue says the machine moved. Decomposed new revenue says which part moved it, and that is where the next dollar goes.
The first version of this chart is usually uncomfortable, because it shows how much of a good year was one bucket carrying the others.
When I broke down the growth at FX Replay, an 18-month run that ended with recurring revenue at 2.6 times its starting point, the contribution came out of email, affiliates, and the shift to annual plans. Paid media took part, though it was never the engine, and knowing that changed which roles got hired next.
Attribution purism kills this metric before it ships. You will never have a flawless assignment of every dollar to a cause, and the decomposition does not need one.
Agree a rule with finance, write it down, and apply it identically each month. Once it holds still, budget conversations get boring in the best way: the bucket that produced the most new revenue for the money it consumed gets funded first, and a bucket flat for three months has to bring a plan rather than a promise.
Where the other twenty-seven metrics live
Cutting the deck to three is a demotion and never a deletion. Cost per click, open rates, trial starts, creative win rates, pipeline, all of it is still measured and still looked at every week, inside the operating review where the team who can move those numbers actually sits.
The distinction that matters is audience. Diagnostic numbers belong to operators who can change them this week. Decision numbers belong to the person allocating capital.
Blur those two audiences and you rebuild the thirty-slide deck within a year, at the cost of the most expensive hour on the company's calendar.
What changes in the room
The first month feels exposed. Three numbers leave nowhere to look away, and when two of them are ugly there is no cost-per-click chart handy to change the subject.
By the second or third meeting the questions change character. Founders stop auditing marketing and start interrogating the business. Why did payback stretch by two months. Is the partner bucket repeatable or was that one deal. If we doubled spend tomorrow, which of these three breaks first.
That shift is the entire return on the discipline. A marketing budget survives contact with a board when the founder can defend it in a single sentence, and a single sentence is only available when there are three numbers to build it from.
Three numbers is how the growth operating system reports upward, and reporting is what decides whether the machine underneath it keeps getting funded.
Build the three-number page
- Put the three on one page: LTGP:CAC, payback in months, and new revenue split into four buckets.
- Settle each calculation with finance once, in writing, so no meeting ever relitigates the arithmetic.
- Set a target and a trigger point beside each number before the month starts, while nobody knows the answer yet.
- Move everything else into the weekly operating review and an appendix you bring to the meeting and rarely open.
- Judge the format on one thing: whether the meeting ends with money moving. If it does not, the three are wrong.
What founders want is to stop feeling responsible for a spend they cannot evaluate. Three numbers, each tied to a decision they are already qualified to make, is what hands that judgment back to them.
If your reporting has grown into a defense document, I build the version that ends in decisions instead. Let's talk.