The SaaS Metrics Investors Ask For Before a Raise

All right, here is the direct answer. Before a raise, investors ask for six numbers: ARR and MRR, net and gross revenue retention, customer acquisition cost, CAC payback, LTV to CAC, and gross margin. Then they ask for the movement schedule underneath, because the schedule is what proves the number.

Here is the part that catches good founders off guard. The list is not the hard bit. Most founders I talk to can recite their ARR to the dollar and have a slide with the retention number on it. What they do not have is the schedule sitting behind it, showing period by period where every dollar came from and where it went. Investors are not really asking what your NRR is. They are asking whether you can show them.

So let me walk you through it the way I would on a call. The six numbers first, because half the arguments in a data room are definitional. Then why the schedule matters more than the headline, the retention pair and how one number hides the other, what a customer costs and how long until you get it back, the two ratios most often computed wrong, the efficiency frame sitting on top, and where founders actually get caught.

The six numbers, defined in one place

Definitions first, because no standard body sets these, and the gap between one founder's definition and one investor's is where deals slow down. Every entry below is written so you can lift it into your own board pack.

1. MRR and ARR. Monthly recurring revenue is the contracted subscription revenue you expect to recognise in a month. ARR is normally that figure times twelve, an annual run rate rather than contracted value or booked revenue. It excludes one-time implementation fees, hardware, pass-through costs and professional services.
2. Gross revenue retention. Take a fixed cohort of customers and measure what is left of their revenue twelve months later after churn and downgrades, counting no upsell at all. Because expansion is excluded, GRR cannot exceed 100 percent. It measures leakage only.
3. Net revenue retention. The same fixed cohort twelve months later, now including expansion, upsell and price increases alongside churn and contraction, and still excluding customers won during the period. Above 100 percent means the existing base grew on its own.
4. Customer acquisition cost. The full cost of winning a customer divided by customers won, including salaries, commissions, advertising, referral fees, credits and discounts. Investors want it split: paid CAC for customers bought through marketing, separate from blended CAC that quietly includes organic and referral wins.
5. CAC payback period. The months of gross-margin-adjusted revenue from a new customer needed to repay the cost of acquiring them. The margin adjustment is not optional, because repaying acquisition cost out of revenue you never keep is not repayment.
6. Gross margin. Revenue less the direct cost of delivering the service: hosting, infrastructure and DevOps staff, application support, and third-party software embedded in the product. What you put in that list moves the number by more than twenty points, which is why investors ask what is in it.

Notice how much judgment sits in those six paragraphs. That is not sloppiness on anyone's part, it is the state of the field, and it is why the schedule matters more than the summary figure.

Why they want the schedule, not the number

Start with the fact nobody says out loud in a pitch. ARR is not an accounting number. ChartMogul's retention research puts it about as bluntly as it can be put: "Annual recurring revenue (ARR) isn't a GAAP metric. Yet it has become the building block of SaaS metrics."¹ Your GAAP revenue is governed by ASC 606, which allocates a contract across its performance obligations and recognises each as it is satisfied.¹⁹ ARR is a forward run rate built from what you invoice. Related cousins, not the same person.

To see how seriously that gap is taken, read what public companies write when a regulator asks. Answering an SEC staff comment on its ARR disclosure, OneSpan told the Commission that "there is no direct relationship between revenue recognized in accordance with ASC 606 and the Company's ARR business metric."² The same filing admits the company holds a customer in ARR for up to ninety days after a contract expires while renewal is negotiated, a defensible judgment that also flatters the number.

i3 Verticals had to spell out the broader point in its own filing: "ARR does not have a standardized definition and is therefore unlikely to be comparable to similarly titled measures presented by other companies."³ The SEC's guidance on key performance indicators asks companies to disclose how a metric is calculated, the estimates behind it, why it is useful, and to say so when the calculation changes.⁴ That is a sound private-company standard too.

The schedule itself is not complicated, and building it is worth more than the artifact. Open with beginning ARR, add new business, add expansion, subtract contraction, subtract churn, land on ending ARR, so the movements reconcile to the balance rather than sitting beside it. ChartMogul's documentation names the categories: "There are six MRR movement types: New Business, Expansion, Contraction, Churn, Reactivation, and Neutral movement."¹⁵ Reactivation and neutral are the two founders leave out, and their absence is usually why a hand-built waterfall refuses to tie.

One purity test before it goes anywhere. SaaS Capital, which lends against recurring revenue and so has money riding on the answer, suggests a ceiling of roughly twenty percent of ARR from impure sources, and is firm that implementation fees do not belong in it: a ninety thousand dollar subscription plus a ten thousand dollar implementation makes year two look like a retention failure when nothing was lost.¹¹ If AI usage revenue or pilots not yet in production sit inside your ARR, decide now how you will describe that.¹

The retention pair, and why one number hides the other

Retention is where I see the most avoidable damage, almost always because only one of the two numbers made the slide. Net revenue retention is the flattering one, because expansion covers for churn. Gross revenue retention is the honest one, because nothing covers for anything. Report only NRR and a careful investor assumes you are hiding GRR, which is a worse position than showing a mediocre GRR.

The picture is sobering, and it moved recently. In the 2026 Aleph and Benchmarkit study, drawing on 342 companies with GRR reported by 226, median gross retention fell to 84 percent for 2025 from 88 the year before, every quartile moving down together. Their framing is the line I would put in front of a board: "At 84%, the typical company loses 16% of its existing ARR every year to churn and contraction."⁶ Median NRR in the same data sat at 102 percent.⁵

Hold that against the benchmark you have been handed. The repeated bar of 110 to 120 percent NRR is real, but it belongs to specific populations rather than private SaaS generally. ICONIQ reports net dollar retention settling in that range across its 127-company dataset, and McKinsey found 130 percent in the top quartile of public SaaS companies against 104 in the bottom.¹⁸²⁶ Private medians run ten to fifteen points below that.

SaaS Capital puts median NRR at 102 percent for companies with contract values of twenty-five to fifty thousand dollars, and warns that "for private SaaS companies, benchmarking retention against public SaaS companies is of limited usefulness."⁸ High Alpha's survey of more than eight hundred companies lands in the same 100 to 104 percent band across ARR tiers.¹⁶ Benchmark against companies your size, or the comparison flatters or frightens you for no reason.

The distance between your two retention numbers is itself a diagnostic. SaaS Capital measured the gap across its survey and its lending file and concluded that "A range of 8-20% is a fairly normal range for a GRR-NRR gap."⁹ Under five points and you are barely expanding, which fewer than one company in ten reports. Over thirty and something one-off is usually in there, a single enormous upsell or a single painful loss, and the investor will find it.

Two mechanical errors to check before publishing either number. The first is letting new customers leak into the expansion line, which Aleph names as the most common: "The most common error is letting new-logo revenue leak into the expansion term."⁵ It inflates NRR and conceals a retention problem. The second is aggregating everything before comparing periods, which lets gains hide losses. SaaS Capital shows a case where annual aggregation reports 87.5 percent GRR while the monthly view has it falling to 50 percent by the fourth quarter.¹¹

Separate logo churn from revenue churn rather than picking one. Counting customers treats a two thousand dollar account and a two hundred thousand dollar account as equals, which overstates the damage because small customers churn most.¹⁰ ChartMogul found companies with NRR below 60 percent running roughly double the customer churn rate of companies at 100 percent or better.¹⁴ Contraction and logo churn also have different cures, one a pricing problem and one an onboarding problem.⁶

What a customer costs, and how long before you get it back

Acquisition cost is the number founders most often present too kindly, usually without meaning to. The a16z partners who wrote the standard piece on misleading startup metrics set the bar plainly: "Customer acquisition cost or CAC should be the full cost of acquiring users, stated on a per user basis."²² Full cost means referral fees, credits, discounts and the loaded cost of the sales team, not the advertising invoice. Expect to be asked for paid and blended separately.

The cost of winning customers has risen, and that is measured rather than felt. Benchmarkit's 2025 report has the new customer CAC ratio up 14 percent in 2024, stated as one sentence: "Companies are spending $2.00 at median to acquire $1.00 of New Customer ARR!"⁷ The fourth quartile spends $2.82. Expansion costs roughly half as much, and Aleph's 2026 read is similar at about eighty cents against a dollar sixty-three.⁷⁵

That single comparison is the strongest argument in finance for spending management attention on the customers you already have. Payback is where I would push back on the figure you have internalised. Twelve months is a rule of thumb, not a measured median, and Benchmarkit says so while noting payback correlates heavily with contract size and that its own median has lengthened 12.5 percent since 2022.⁷

The measured figures run longer than the convention. ICONIQ has top-quartile payback at 15.9 months in 2025, out from 12.3 in the 2020 to 2021 window and peaking at 17.3 in 2024.¹⁸ McKinsey, across 100 public SaaS companies above one hundred million dollars of revenue, found top-quartile median payback of 16 months and bottom-quartile 47.²⁶ If your payback is fourteen months, you are not behind.

Segment matters more than any single benchmark, and Bessemer's portfolio guidance is the most usable version I know: under twelve months selling to small business, under eighteen to mid-market, under twenty-four to enterprise, because enterprise customers stay longer and are worth more.²⁷ Their data also shows payback lengthening as a company scales, since your earliest adopters were the cheapest customers you will ever get. If you cannot say by how much or why, that is the problem.

The two ratios that get computed wrong most often

LTV to CAC first. The three-to-one convention traces to David Skok, and his wording is more demanding than the version in circulation: "The best SaaS businesses have a LTV to CAC ratio that is higher than 3, sometimes as high as 7 or 8."²⁰ Three is a floor, not a target, and Bessemer independently recommends investing in acquisition once the ratio clears three.²⁷ His companion guideline is tighter than most people quote too: the best businesses recover acquisition cost in five to seven months.²⁰

Now the honest caveat, because I would rather you hear it from me than from a partner in a diligence meeting. Lifetime value is the softest number on the list. Bill Gurley's critique is still the best written: the variables inside the formula pull against each other rather than moving independently, so raising average revenue tends to raise churn, and spending more on marketing raises cost and lowers customer quality.²¹

Gurley also names the abuse of dividing total spend by total customers, which quietly credits organic wins to the paid channel. a16z cites him approvingly and adds the error I see most often, computing lifetime value on revenue or gross margin instead of net profit.²² If your data is thin, a twelve or twenty-four month historical figure is more defensible than a projected lifetime, and easier to support when someone asks how you built it.

Gross margin is the second, and here the best sources openly disagree. a16z holds that "software companies should have very high gross margins, in the 80%-90% range."²³ Bessemer, looking at its own cloud portfolio, reports that "the average gross margin for a cloud business regardless of maturity is 65-70%."²⁷ Benchmarkit's survey median is 77 percent on total revenue and 81 on subscription.⁷ Those are not contradictory findings. They are different definitions of cost of sales.

Definitions vary because no standard forces them to converge. SaaS Capital says it plainly: "Surprisingly, GAAP does not clearly define what should be included in a SaaS company's Cost of Sales."¹² Their recommended contents are hosting, infrastructure labour, application support, embedded third-party software and other direct delivery cost, leaving out sales commissions, capitalised development amortisation and product management. On that basis they put licence gross margin at 80 to 85 percent.

The practical consequence is the one they flag: without knowing the margin on subscription revenue specifically, nobody can compute your payback correctly, yours included. So decide what sits in cost of sales, write it down, and apply it to every period in the schedule rather than only the current one. A definition you can state is worth more than a number you cannot defend.

The efficiency frame that now sits on top of all of it

Around 2022 the question investors were asking changed, and the metrics did not change with it so much as get reweighted. KeyBanc's sixteenth annual private SaaS survey describes the shift in the companies' own language, saying they "continue to shift their priorities from a growth-at-all-costs strategy to one of balanced growth and profitability."¹⁷ That release is a headline summary with no disclosed sample size, so treat it as direction rather than measurement.

ICONIQ quantifies it better than anyone I have read, because it prints the sample size on every chart. Across its 127-company dataset, median Rule of 40 performance fell from 78 percent before 2020 to 47 in 2023 and 2024, recovering to about 50 in 2025. More telling is the composition: free cash flow's share of that composite rose from 20 percent before 2020 to 44 in 2025, while growth's share fell from 80 to 56.¹⁸ The scoreboard did not just get harder. It started measuring something else.

The gap between funded and self-funded companies is worth seeing too. SaaS Capital's 2026 survey, with more than a thousand private SaaS respondents, found 83 percent of bootstrapped companies within two points of breakeven or profitable, against 52 percent of equity-backed ones, while bootstrapped companies grew 20 percent at median against 25.¹³ That is the whole efficiency argument in one comparison: a five point growth difference bought with a great deal more spending.

Two cautions on the frames, since both get quoted loosely. Brad Feld, who popularised the Rule of 40 and is candid that he heard it at a board meeting rather than inventing it, attached a scale condition that almost always gets dropped: these are for SaaS companies at scale, "assume at least $50 million in revenue."²⁵ Applying it to a company at four million dollars of ARR is a category error. It is also hard to hit: McKinsey found that "businesses exceeded Rule of 40 performance only 16 percent of the time."²⁶

Burn multiple is the other frame in every deck now, useful precisely because it refuses to isolate anything. David Sacks, who coined it at Craft Ventures in 2020, describes it simply: "The higher the Burn Multiple, the more the startup is burning to achieve each unit of growth."²⁴ A margin problem, a churn problem and a sales efficiency problem all present in that one ratio, which is why an investor can start there and work outward.

Worth knowing what McKinsey found about the metrics that do not predict value, though. Of roughly twenty it tested, only ARR growth, net retention, payback period and free cash flow margin correlated strongly with revenue multiples, and it found almost no correlation for the magic number or ARR per employee.²⁶ Not every ratio in circulation earns its slide, and a deck full of them reads as noise rather than command.

Where founders actually get caught

This is the part I most want you to take away, because it is where the money moves. Founders rarely lose value on the level of a metric. They lose it on whether the numbers hold up when someone competent pulls on them. Grant Thornton's first quarter 2026 survey of 185 US respondents at companies between one hundred million and four billion dollars of revenue found the quality or completeness of target financials to be the top barrier to completing deals, cited by 52 percent.²⁸

That ranked ahead of valuation gaps at 46 percent and diligence complexity at 42. And what it looks like in practice is not a rejection but a slowdown, which is expensive on its own. As Palash Misra of Grant Thornton describes it, "When financials are incomplete or difficult to validate, we tend to see more staging in the process."²⁸ Buyers add checkpoints while your exclusivity window runs and your numbers age.

The instrument doing the pulling is usually a quality of earnings review, and it is worth knowing its shape before you meet one. Baker Tilly states that "A QOE typically takes four to six weeks to complete."²⁹ Warren Averett puts middle-market work slightly tighter: "The majority of middle market analyses typically take three to four weeks."³⁰ Both name the same two adjustments that hit software companies hardest, cash to accrual and revenue recognition.

Those are not filing errors. They are adjustments to the earnings figure your valuation is a multiple of, which is why they are worth fixing before anyone is looking. An audit does not spare you them either. Baker Tilly makes the point that items entirely correct under GAAP can still create deal issues, because an audit opines annually while a buyer prices off the trailing twelve months and reads the numbers monthly.²⁹

Armanino, writing from the buy-side, is blunt about where processes stall: of everything in a diligence workstream, the quality of earnings report with its adjusted earnings figure is the one where "It is the number one area where diligence tends to bog down."³¹ Their named causes are mundane and avoidable, including duplicate requests across two teams and analysts chasing items under ten thousand dollars.

So what prevents all this. Get onto accrual accounting before you need it, which SaaS Capital pins to a threshold worth writing on the wall: by the time a company "approaches $5 M in ARR or seeks outside institutional funding, it's time to be moving to accrual accounting."¹¹ Write down your ARR definition, your cost of sales definition and your retention method, and apply them to every period. Build the movement schedule monthly as a habit, not as a diligence deliverable.

On audits, calibrate rather than panic. Kruze Consulting, working from a book of venture-backed startups, reports that "venture capitalists don't usually start asking for an audit until a company is Series C or later," at observed costs of twenty to fifty thousand dollars and two to four months.³² Before that, clean accrual books and a schedule that ties will carry you further than an expensive engagement nobody asked for.

The pattern I have watched for twenty-three years is that the founder who wins the diligence conversation is not the one with the best numbers. It is the one who knows where each number came from, says the unflattering part first, and has the schedule ready before it is asked for. Investors are forming a view about whether you know your own business, and a number you can decompose answers that question in a way a number you can only assert never will.