Before you treat 70% as a success story
SaaS Gross Margin: the delivery-cost tripwire that caps how profitable you can get
A slide that prints “gross margin is 72%” without the COGS buckets, the revenue basis, or what you will do if it misses is not validation. SaaS / cloud gross margin asks a smaller, honester question: after hosting, implementation, and the delivery or support costs that belong in COGS, what share of each revenue dollar is left to fund S&M, R&D, and eventually profit — against a bar you wrote first? Matching it is a decision signal — not a forecast that you will succeed, and not a substitute for naming the weakest demand or retention assumption first.
Last updated: October 20, 2026
Direct answer
What is SaaS gross margin — and why write the tripwire first?
Gross margin = (Revenue − Cost of Goods Sold) ÷ Revenue. For cloud / SaaS it answers: after hosting, implementation, and delivery or support costs that belong in COGS, what share of each dollar is left to fund growth and eventually profit. Matching or missing a pre-written tripwire is a decision signal — not a prediction that you will succeed. Bessemer reports this on company financials. Do not invent an ARR-only formula and call it the same sitting.
Key takeaways
What to remember
- Gross margin = (Revenue − COGS) ÷ Revenue. Gross profit is Revenue − COGS. Bessemer measures how effectively you deliver software: a high margin leaves more dollars for operating expenses and growth. Always show Revenue, the COGS buckets, and the % — not only the %. Write whether the revenue basis is GAAP period revenue or something else. Do not mix ARR into a GAAP gross-margin sitting without labeling it.
- Bessemer’s Scaling to $100 Million (21 September 2021): investors expect a tight band. Average cloud gross margin is 65–70% regardless of maturity; the middle 50% stays within ~60–80%. Average GM by ARR band: $1–10MM 70%; $10–25MM 70%; $25–50MM 65%; $50–100MM 65%; $100MM+ 70%. Median GM by ARR band: 72% / 70% / 65% / 70% / 70%. Top-performing floors: 85%+ / 80% / 75%+ / 80%+ / 80%+. Typical cloud COGS: hosting, software implementation, and services including customer success when classified in COGS.
- Bessemer’s CEO tactical guide (30 March 2023): median gross margin for high-growth public cloud companies is 77%. Holding other factors constant, an 80% GM business at 40% profit versus the same business at 60% GM → only 20% profit; that 80% vs 60% gap cuts business value by at least 50%. Improving GM rarely trades off growth investment the same way cutting S&M does — opex sits below gross profit.
- Honesty caveat, required: Bessemer notes some of the strongest portfolio companies sit below the 60–80% band when COGS structure demands it. Twilio ran ~50% GM for much of its life because telecom pass-through sat in COGS. Low GM is not automatic failure; it forces a different unit-economics and pricing tripwire. Do not treat ≥70% or ≥77% as proof of success.
- Write the tripwire before you measure: signal (gross margin %) + threshold (e.g. aim at the 65–70% average band, or “never below 60% without a written COGS reason”) + definition (what is in COGS vs opex; GAAP period; include/exclude professional services) + period + named action (fix hosting / implementation classification / price services / keep experimenting / do not scale spend that worsens delivery cost). A Yibud score is not gross margin and not a success prediction.
Why it matters
Why a 70% slide can still cap how much you can earn
Founders treat “we are a software company, so margin is fine” as a feeling. Feelings do not name whether this period’s delivery costs left enough gross profit to fund growth and, later, profit. Two Bessemer sittings are useful here as instruments, not as personal odds: Scaling to $100 Million on the tight band, and the CEO tactical guide on the valuation ceiling.
Bessemer Scaling: delivery effectiveness, not a slogan
Bessemer writes that gross margin measures how effectively you deliver software. High GM means more dollars available for operating expenses and growth rather than product delivery. Investors expect a tight band — average 65–70% regardless of maturity, middle 50% ~60–80% — because software’s marginal replication cost should be low. That is their portfolio sitting, not a proof you will succeed.
CEO guide: 80% vs 60% is a valuation ceiling, not a vibe
The 30 March 2023 CEO guide is blunt: gross margin acts as the limit to ultimate profitability and has an enormous impact on valuation. Median GM for high-growth public cloud is 77%. Hold other factors constant: 80% GM at 40% profit versus 60% GM at 20% profit. That gap cuts business value by at least 50%. Improving GM rarely costs growth investment the way cutting S&M does, because R&D, S&M, and G&A sit below gross profit.
Twilio ~50%: a structural low GM is not “failed SaaS”
Bessemer’s honesty line: some of the strongest cloud companies in the portfolio sat below the 60–80% band. Twilio’s GM was ~50% for much of its life after the 2009 seed, because it paid telecom providers in COGS. Low GM is a different tripwire — pricing, mix, and unit economics have to carry a heavier delivery cost — not an automatic kill. Do not treat ≥70% as a health certificate, and do not treat a Twilio-like structure as shame.
The six terms
Revenue, COGS, gross profit, gross margin, opex, tripwire
These six fields are the instrument. If any one is fuzzy, the number is a slide, not a tripwire. This page does not publish a grade table beyond Bessemer’s published averages, medians, top-performing floors, the 65–70% / ~60–80% band, Twilio’s ~50% caveat, and the CEO-guide 77% / 80-vs-60 vignette.
Revenue — name the period basis; do not invent an ARR-only GM
The period revenue used in the margin. Bessemer reports gross margin on company financials — revenue after COGS, divided by revenue. Founders often mix trailing ARR into that sitting. If you use ARR, say so; it is a different basis. This page does not invent an ARR-only gross-margin formula. Write GAAP period vs ARR before the first report, and do not switch after the close.
COGS for cloud — hosting, implementation, services including CS
Bessemer’s cloud COGS list: hosting costs, software implementation costs, and services costs, including customer success when those costs are classified in COGS — all treated as variable delivery costs. This page does not invent a GAAP rulebook. Write which buckets you include, and whether professional services sit in COGS or below. Reclassifying CS into opex after the close is how a 65% becomes a kinder 78%.
Gross profit — Revenue minus COGS
The dollars left after delivery. Bessemer: a company’s revenue after the cost of goods sold. This is the pool that can fund operating expenses and residual cash flow. A 72% slide that hides $18,000 of implementation is a different pool from a 72% that lists it.
Gross margin — gross profit ÷ Revenue
The share of each revenue dollar left after COGS. Bessemer: “a company’s revenue after the cost of goods sold (or gross profit) divided by revenue.” It is a percent, usually discussed in the 60–80% cloud band. It is not Rule of 40 (growth% + profit%). It is not LTV:CAC. It is not NRR%.
Operating expenses sit below gross profit
After you deliver the product and incur COGS, Bessemer writes, you can spend the remaining gross profit on R&D, S&M, and G&A, and retain any residual as cash flow. The CEO guide’s useful framing: improving GM rarely comes at the expense of investment in growth the way cutting S&M does. Do not “save” margin by starving the opex that sits below it and calling that delivery excellence.
The tripwire — written before you measure
Signal + threshold + definition (COGS vs opex; GAAP period; include/exclude professional services) + period + named action (fix hosting or implementation classification / price services / keep experimenting / do not scale spend that worsens delivery cost). Write it on a dated page before the first report. A kinder bar after the results is a recap, not a tripwire. Matching the bar is a decision aid. It does not prove you will succeed.
The arithmetic
Delivery-cost math — not a proprietary Yibud formula
There is no Yibud gross-margin formula. Use the sitting Bessemer actually wrote on company financials, name the COGS buckets, and do not swap ARR for GAAP revenue after the report. The fictional worked example later on this page is invented so you can see the crossing; it is not a benchmark.
Gross margin (Bessemer wording)
Gross profit = Revenue − COGS. Gross margin = Gross profit / Revenue.
Same period on every term. Bessemer: revenue after the cost of goods sold, divided by revenue. If you drop implementation or CS-in-COGS after the close, you changed the definition. Do not invent a “Yibud adjusted GM” that quietly parks delivery people in opex.
CEO-guide ceiling — 80% vs 60%, their numbers
80% GM business at 40% profit vs the same business at 60% GM → 20% profit (other factors held constant). That gap cuts business value by at least 50%.
This is Bessemer’s 30 March 2023 vignette, not a Yibud identity. It is why a delivery-cost leak is a valuation ceiling, not a rounding error. Use it as vocabulary for the profit room GM leaves. Do not treat 80% or 40% as a seed-stage target you must print this quarter.
Period crossing (fictional inputs)
Q1 GAAP revenue $80,000. Hosting $8,000. Implementation $6,400. CS classified in COGS $9,600. COGS = $24,000. Gross profit = $56,000. Gross margin = 56,000 / 80,000 = 70%.
These dollars are invented so the arithmetic is visible. They are not Bessemer’s portfolio averages, not the 77% public-cloud median, and not a Yibud report. The 70% would clear a ≥65–70% bar you wrote. A slide that dropped implementation and CS-in-COGS would have printed 90% on hosting alone. Same company. Different definition. Different decision.
Do not mix a GAAP numerator with an ARR denominator. Do not annualize a 30-day margin by multiplying. Do not hide professional services after the close. Bessemer’s list is hosting, implementation, and services including CS when classified in COGS — write your own buckets before the first report. The tripwire names the period and the buckets first.
Published bands
Bessemer averages, medians, and top floors — labeled, not a grade table
The rows below are what Scaling to $100 Million actually named, by ARR band. They are not a Yibud letter-grade chart, not a personal odds table, and not a Cloud 100 LinkedIn median. Use them as vocabulary, then write your own tripwire for the period you can actually measure.
| ARR band | Average GM | Median GM | Top-performing floor |
|---|---|---|---|
| $1–10MM | 70% | 72% | 85%+ |
| $10–25MM | 70% | 70% | 80% |
| $25–50MM | 65% | 65% | 75%+ |
| $50–100MM | 65% | 70% | 80%+ |
| $100MM+ | 70% | 70% | 80%+ |
Source: Bessemer Venture Partners, “Scaling to $100 Million” (Mary D’Onofrio, Ethan Ding & Atlas Editors; 21 September 2021). Averages from the gross-margin section; medians from the published ARR-band medians; top floors from Scaling Lesson 4. Labeled as their private-cloud portfolio sitting, not a law for every seed app.
The tight band — 65–70% average, middle 50% ~60–80%
Bessemer: it is perhaps the only operating or cost metric with very little wiggle room. Average cloud GM regardless of maturity is 65–70%; the middle 50% stays within ~60–80%. Lesson 2’s aim is 65–70% over time. That is their investor expectation, not proof a 68% company will succeed.
Twilio ~50% — strongest companies can sit below the band
Required caveat. Bessemer: some of the strongest portfolio companies had GM below that range. Twilio’s GM was ~50% for much of its life after the 2009 seed, because telecom service providers sat in COGS. Low GM is not automatic failure. It forces a written COGS reason and a different pricing / unit-economics tripwire.
CEO guide — 77% median for high-growth public cloud
Bessemer, 30 March 2023: the median gross margin for high-growth public cloud companies is 77%. That is a public-cloud median on that page, not a seed law, and not a Yibud grade. Do not flatten it onto a fourteen-person company and call the debate settled.
80% vs 60% — at least a 50% valuation cut, their vignette
Same CEO guide: two identical software businesses; 80% GM at 40% profit versus 60% GM at 20% profit, other factors constant. That gap cuts business value by at least 50%. Use it to see why delivery cost is a ceiling on Rule of 40 profit, not as a homemade “80 or die” table.
One-line opex context only — this page is not an opex guide. Bessemer: by $100MM+ ARR, R&D averages 35% of revenue; S&M still represents over 50% of revenue even at maturity; G&A averages 20% by $50MM+.
Scope notes, still the sources: Scaling to $100 Million is Bessemer’s private cloud portfolio sitting (published 21 September 2021). The CEO tactical guide is a 30 March 2023 profitability essay. Do not invent First Round / KeyBanc / a16z / Cloud 100 LinkedIn GM tables on this page. Do not treat ≥70% or ≥77% as proof you will succeed.
How to run it
How to run a gross-margin tripwire without faking scale
This is Yibud’s suggested sequence, not a Bessemer checklist and not a GAAP audit. Honor the tripwire you wrote before the first period report. If you cannot name the decision the number would change, you are building a slide.
- 1
Write the tripwire before you look
Signal is gross margin % for a named period, plus Revenue and the COGS buckets. Threshold is the miss / pass line you will honor — Bessemer’s 65–70% average band and “never below 60% without a written COGS reason” are classic gates, not a seed law, and not proof you will succeed. Name GAAP vs ARR, which buckets sit in COGS, and whether professional services are in. Action is fix hosting or implementation, price services, keep experimenting, or do not scale spend that worsens delivery cost. Do not rewrite the bar after the report.
- 2
Lock the period and the buckets
Same window on Revenue and every COGS line. If you only have 30 days, report a 30-day GM and say so — do not multiply by twelve and call it annual. Mixing a quarterly implementation dump into a monthly hosting number is how a leak prints 81%.
- 3
Compute GM. Put Revenue and the COGS buckets on the dated page.
Subtract COGS from Revenue. Divide. Do not publish 72% and hide $9,600 of CS-in-COGS behind “software is 90%.” The buckets are the sitting; the percent is the compression. Optional actions from the CEO guide — hosting negotiation / FinOps, charge for implementation or move it to a third party, support tiers — are examples of named actions, not new numeric benchmarks.
- 4
Read neighbors as separate sittings. Do not average them into GM.
Rule of 40 uses profit after this ceiling. LTV:CAC and CAC payback often use a GM-adjusted contribution. Magic Number, burn multiple, NRR, and Quick Ratio answer different decisions. Do not average GM with any of them. Use those pages for those instruments.
- 5
Honor the action. If you are pre-revenue, use a different instrument.
If the period closes and you missed the bar you wrote, do the thing you named — fix delivery cost, price services, keep experimenting, or do not scale spend. Do not reclassify COGS into opex so the % clears 70%. Pre-revenue, pure marketplace take-rate confusion, or hardware-heavy COGS: use willingness-to-pay, smoke, or interviews first. A Yibud score is not gross margin.
What you leave with
The gross-margin tripwire (copy this)
If the sitting produced a percent and no dated action, you did not finish. Fill the brackets. Leave nothing as “margin is fine,” “soon,” or “we’ll know it when the round happens.”
The fill-in contract
Product: [name]. Period: [month or quarter — books close date], written before the first report. Hypothesis this number can falsify: [after delivery COGS, enough of each revenue dollar is left to fund growth / we can scale spend without worsening delivery cost]. Revenue basis: [GAAP period revenue — not unlabeled ARR]. COGS buckets: [hosting / implementation / services including CS — list which are in]. Professional services: [in COGS / excluded]. Revenue: [ ]. COGS: [ ]. Gross profit: [ ]. Gross margin: [ ]. Signal: [GM% and the buckets, and/or change vs last period]. Threshold: [e.g. honor a miss if quarterly GAAP GM is below 65%, or if it falls below 60% without a written COGS reason — Yibud template using Bessemer’s 65–70% average band and ~60–80% middle-50% floor, not a homemade letter grade, and not the 77% public-cloud median treated as a seed target]. Action if miss: [fix hosting / price implementation / keep experimenting / do not 3× spend that worsens delivery cost]. Written COGS reason if below 60%: [Twilio-like pass-through / hardware / other — or none]. Compare the next period on the same definitions. A Yibud score does not fire this line.
The “below 65% / below 60% without a written reason” line is Yibud’s suggested template from Bessemer’s 65–70% average and ~60–80% middle band — not a census, not permission to skip the COGS buckets, and not permission to treat ≥70% as success. Rewrite the threshold before you measure if you have a reason. Do not rewrite it after.
Example tripwire (Yibud’s suggested template, not a law)
If, for the calendar quarter that closes on the last day, GAAP gross margin (Revenue − hosting − implementation − CS-in-COGS) / Revenue is below 65%, or if it is below 60% and we have no written COGS reason (Twilio-style pass-through or equivalent) → we do not 3× paid acquisition next quarter; we fix delivery cost first and name the weakest pricing or COGS assumption. Do not drop implementation after the close. Do not mix ARR into the denominator. Report Revenue and the buckets next to the %.
A 70% GM on one quarter is still one period. A Yibud report score does not fire this tripwire. If you have no revenue yet, this template is the wrong first instrument — use willingness-to-pay, a smoke test, or interviews instead.
The fork
When to use a gross-margin tripwire — and when not to
Gross margin is the wrong first instrument when there is no revenue yet, when a marketplace take-rate is being confused with software GM, or when hardware-heavy COGS makes the cloud 65–70% band the wrong vocabulary. It is the right instrument when a period of software / cloud revenue can close, COGS can be named, and you are deciding whether delivery cost leaves enough room before you scale spend.
Use it when you have software/cloud revenue, named COGS, and a spend-or-fix decision
You can name period Revenue and the delivery buckets. You are about to 3× a channel, tell yourself “GM is 70% so we are fine,” or skip the COGS list because “we are software.” That is when a dated % plus buckets earns its keep. Optional: run Rule of 40, LTV:CAC, CAC payback, Magic Number, burn multiple, NRR, or Quick Ratio as separate sittings — do not average them into a homemade “efficiency score.”
Do not start here when you are pre-revenue, confusing take-rate with GM, or still need first money, reach, or a who-filter
No paying customers yet? There is no honest Revenue or COGS period. Do not invent a forecast numerator. Pure marketplace? Take-rate is not SaaS gross margin — do not paste Bessemer’s 65–70% cloud band onto GMV. Hardware-heavy COGS, or a services book without software delivery? The cloud tight band is the wrong first vocabulary; write a different COGS reason or use another instrument. Still writing a deposit / pre-order / priced checkout? That is willingness to pay. Still testing whether anyone clicks? Smoke, fake-door, or landing-page tests. Still asking whether current users would be very disappointed? Ellis PMF survey. Still writing Stop / Pivot / Continue? Those are the decision contracts. This page supplies one delivery-cost instrument those contracts can later read. It does not replace them. A Yibud score is not gross margin.
Worked example
One quarter you can copy (illustrative — fictional)
The quarter below is illustrative. Names, product, and dollars are invented so you can see the arithmetic — including a 90% slide that dropped two COGS buckets. It is not a study, not an industry benchmark, and not a Yibud report. Copy the method, not the numbers.
The promise, written before the quarter closes
Illustrative names: founder Priya Shah at Inkline, contract-notes software for boutique law firms. Recurring software revenue exists. A GAAP quarter can close. The tempting trophy is a board slide that prints GM = 90% from hosting alone, no implementation line, no CS-in-COGS, and ARR in the denominator. The tripwire is written first, on 3 January: honor a miss if Q1 GAAP GM is below 65%, or below 60% without a written COGS reason — Yibud’s suggested template from Bessemer’s 65–70% average and ~60–80% middle band, not a homemade letter grade, and not permission to skip the buckets.
The inputs, kept honest
Q1 GAAP revenue: $80,000. Hosting $8,000. Implementation $6,400. CS classified in COGS $9,600. COGS = $24,000. Gross profit = $56,000. Gross margin = 70%. A slide that dropped implementation and CS-in-COGS would have printed 90% ($80,000 − $8,000) / $80,000. Same company. Different definition. Different decision.
The tripwire, written 3 January
Signal: Q1 GAAP gross margin and the three COGS buckets. Threshold: honor a miss if GM is below 65%, or below 60% with no written COGS reason, before we 3× paid acquisition. Window: books close 31 March. Action if miss: do not 3× paid in Q2; fix delivery cost first; name the weakest pricing or COGS assumption. A Yibud score does not fire this line. The 65 / 60 pair is Yibud’s suggested template from Bessemer — not a census.
What a miss looks like — and what must not happen
Illustrative close: 70% clears the ≥65% bar they wrote. The 90% hosting-only sketch is the slide they refused to swap in after the close. Had GM printed 58% with the same $80,000 revenue and $33,600 COGS, that would have been a miss unless they had a written Twilio-like reason. Do not mix ARR ($120,000 trailing) into the denominator after the close and call 53% a different company. Do not treat 70% as proof the startup will succeed. Keep the next quarter on the same buckets.
Write the tripwire before the first report. If the period closes and GM misses the bar you wrote — or only “clears” it after you drop a COGS bucket — honor the action. If the % and the buckets both cleared the bar, continue to the next cheapest test — usually the next period on the same definitions, or a kill/pivot contract that now has a number to read. The method is the dated page, not the 72% slide. Neither 70% nor a Yibud score predicts success.
Not the same page
Gross margin vs Rule of 40, LTV:CAC, CAC payback, Magic Number, burn multiple, NRR, Quick Ratio, vanity metrics, WTP, Ellis, failure rates, Yibud scores, and decision contracts
Several Yibud pages sit next to this one. Mixing them up turns a growth-plus-profit sum, a lifetime ratio, months-to-recover-CAC, a sales-efficiency multiple, a cash burn multiple, a cohort retention %, a same-period inflow÷outflow, a waitlist, or a mid-band engine number into “GM is 70%, so scale.”
Rule of 40 — growth% + profit%, not this delivery-cost ceiling
That page asks whether revenue growth rate plus profit margin (or FCF) cleared a ≥40 bar. Gross margin is an input ceiling on how high profit can go — the CEO-guide 80% vs 60% vignette is why. A 42 Rule of 40 does not tell you GM is 70%. A 70% GM does not tell you growth plus EBITDA is healthy.
SaaS Rule of 40 →LTV:CAC — lifetime value ÷ CAC, often GM-adjusted, not this %
That page asks whether a cohort’s lifetime (or 12/24-month) contribution cleared a multiple of CAC. The numerator is often gross-margin-adjusted. This page owns the GM math itself. A 3.2 LTV:CAC does not tell you period GM. A 70% GM does not tell you lifetime value is three times CAC.
LTV:CAC for startups →CAC payback — months to recover CAC on contribution, not this %
That page asks how many months of contribution (or GM-adjusted ARR) recover fully-loaded CAC. This page is period delivery cost as a percent of revenue. A 7-month payback does not tell you GM is 70%. A 70% GM does not tell you LinkedIn ads pay back in months.
CAC payback for startups →SaaS Magic Number — S&M → incremental revenue, not delivery cost
That page asks how efficiently prior-period sales and marketing spend turned into incremental subscription revenue. This page is Revenue minus COGS over Revenue. A 1.2 Magic Number does not tell you GM. A 70% GM does not tell you last quarter’s S&M paid for itself in new ARR.
SaaS Magic Number →Burn multiple — company cash vs net new ARR, not this %
That page asks how much company cash you burned per dollar of net new ARR. This page is not cash. A 2× burn multiple does not tell you GM. A 70% GM does not tell you the company is burning too much cash per new ARR.
Burn multiple for startups →NRR — cohort retention of a starting base, not delivery cost
That page asks whether customers live at period start paid you more or less after churn, downsell, and expansion (NRR% / GRR%). This page asks what share of this period’s revenue survived COGS. A 110% NRR does not tell you GM is 70%. A 70% GM does not tell you the starting cohort retained 110%.
SaaS NRR →Quick Ratio — same-period inflow ÷ outflow, not this %
That page asks whether this period’s new + expansion outran churned + contraction (a multiple). This page is GM%. A 4× Quick Ratio does not tell you delivery cost left 70 cents. A 70% GM does not tell you the bucket is not leaking.
SaaS Quick Ratio →Vanity metrics — hygiene class, not this delivery-cost tripwire
That page is the sorting hat: hits, cumulative signups, and waitlist size feel like progress and do not tell you what to do next. This page is one specific actionable percent. A pretty GM chart with implementation dropped and no tripwire can still be vanity. A waitlist is not Revenue.
Vanity metrics for startups →Willingness to pay — first money on a deal, not period GM
That page is a deposit, pre-order, paid concierge, or priced checkout as proof that money moved once. Gross margin needs a revenue period and named COGS. A paid first month with no books should not be running this page yet.
Willingness-to-pay test →Sean Ellis PMF survey — disappointment among current users
That page’s signal is the percent of recent real users who would be very disappointed if the product disappeared. Bessemer’s 65–70% band is a delivery-cost sitting, not Ellis’s question. Do not treat “very disappointed” as gross margin, and do not treat 70% GM as Ellis’s bar.
Product-market fit survey →Startup failure rate — population frequencies, not the math
That page is what the failure data says at population level — BLS establishment survival, venture-backed “failure” as reported, and why a base rate is not your fate. This page teaches one delivery-cost instrument those tables do not compute. It does not invent a “percent of startups that die from sub-65% GM.”
What percentage of startups fail →Yibud scores — not gross margin, and not a success prediction
The validation-score and score-calculator pages explain Startup MRI numbers from a deterministic rule engine on a one-line idea plus five questions. Those scores point at the weakest assumption to test next. They are not gross margin. They are not Rule of 40, LTV:CAC, CAC payback, Magic Number, burn multiple, NRR, or Quick Ratio. Do not read a mid-band monetization score as “GM is 70%,” and do not read 70% GM as a Yibud GO.
Kill, pivot, critical assumption, decision — contracts that can read this number
Those pages write Stop, rewrite one dimension, name the one claim that would kill the idea if false, or decide after evidence. This page can supply the delivery-cost instrument those contracts later read: a dated GM% and COGS buckets against a bar you wrote first. A 72% slide with no implementation line cannot fire a kill line honestly.
Where Yibud fits
Neither gross margin nor a Yibud score predicts success
Yibud is a free, no-signup startup idea validator. Scores come from a deterministic rule engine, not from a language model guessing success. Optional AI text, when it is used, only polishes prose. It does not invent the number. A Startup MRI report can name a weak monetization or distribution assumption, then a Day-7 Continue / Refine / Re-test / Stop signal. The score is a flashlight on the weakest assumption to test next — not gross margin, not Rule of 40, not LTV:CAC, not CAC payback, not Magic Number, not burn multiple, not NRR, not Quick Ratio, not a forecast. Gross-margin work sits next to that assumption: write the tripwire, do not hope a 72% slide means delivery cost leaves room. If you already have a report, start with the named assumption, not the overall score. Then run the period yourself. The method on this page stands alone if you never open the analyzer.
Analyze my idea →Common mistakes
What usually wastes a gross-margin week
Reporting GM without listing the COGS buckets
Bessemer’s cloud list is hosting, implementation, and services including CS when classified in COGS. A 72% with $18,000 of hidden implementation is a different company from a 72% with $1,800. Put Revenue and the buckets on the dated page.
Mixing ARR into a GAAP GM sitting without labeling it
Bessemer reports gross margin on company financials. Trailing ARR in the denominator is a different basis. Write GAAP period vs ARR before the first report. Switching after the close is how a 70% becomes a slogan.
Treating a Twilio-like structural low GM as “failed SaaS”
Bessemer: some of the strongest portfolio companies sat below the 60–80% band. Twilio ran ~50% GM for much of its life because telecom pass-through sat in COGS. Write the COGS reason. Change the pricing / unit-economics tripwire. Do not treat low GM as automatic failure, and do not treat ≥70% as a health certificate.
Scaling S&M before the GM tripwire when delivery costs are the leak
The CEO guide: GM is the limit to ultimate profitability; improving it rarely trades off growth investment the way cutting S&M does. If hosting and implementation are the leak, 3× paid acquisition pours spend into a thinner and thinner gross-profit pool. Write the bar before you scale.
Treating ≥70% or ≥77% as proof you will succeed
Bessemer’s average band is 65–70%. The CEO-guide public-cloud median is 77%. Clearing the bar you wrote means that period cleared a gate — not that the startup will succeed. It is not Ellis’s very-disappointed question, not a deposit, and not a Yibud GO. Neither gross margin nor a Yibud report score predicts success.
Sources
Where these ideas come from
- Mary D’Onofrio, Ethan Ding & Atlas Editors / Bessemer Venture Partners, “Scaling to $100 Million,” 21 September 2021 — Used for Lesson 2’s 65–70% aim; average cloud GM 65–70% regardless of maturity; middle 50% ~60–80%; COGS examples (hosting, implementation, services including CS); average GM by ARR band 70 / 70 / 65 / 65 / 70; median GM by ARR band 72 / 70 / 65 / 70 / 70; top-performing floors 85%+ / 80% / 75%+ / 80%+ / 80%+; Twilio ~50% GM caveat; one-line opex context at scale (R&D 35% by $100MM+ ARR; S&M over 50% even at maturity; G&A 20% by $50MM+). Do not invent Cloud 100 LinkedIn medians or extra grade tables.
- Brian Feinstein, Caty Rea, Janelle Teng Wade & Sameer Dholakia / Bessemer Venture Partners, “A CEO’s tactical guide to driving profitable growth,” 30 March 2023 — Used for median GM of high-growth public cloud = 77%; 80% GM → 40% profit vs 60% GM → 20% profit (other factors constant); ≥50% valuation cut from that gap; improving GM rarely trades off growth investment the same way cutting S&M does; optional tactical examples (hosting negotiation / FinOps, charge for implementation or third-party PS, support tiers) as actions, not new numeric benchmarks. This page does not invent Gainsight ARR-per-CSM numbers.
In one paragraph
Summary you can quote
SaaS / cloud gross margin = (Revenue − COGS) ÷ Revenue. After hosting, implementation, and delivery or support costs that belong in COGS, it asks what share of each dollar is left to fund growth and eventually profit. Bessemer Scaling (21 September 2021): average cloud GM 65–70% regardless of maturity; middle 50% ~60–80%; averages by ARR band 70 / 70 / 65 / 65 / 70; medians 72 / 70 / 65 / 70 / 70; top floors 85%+ / 80% / 75%+ / 80%+ / 80%+; Twilio ~50% GM when telecom sat in COGS. CEO tactical guide (30 March 2023): high-growth public-cloud median 77%; 80% GM at 40% profit vs 60% GM at 20% profit cuts value by at least 50%. Write signal + threshold + definition + period + action before you measure; report Revenue and the COGS buckets, not only the %. Matching the gate is a decision signal. Neither gross margin nor a Yibud score predicts success.
FAQ
Questions founders actually ask
What is SaaS gross margin?
(Revenue − Cost of Goods Sold) ÷ Revenue. For cloud / SaaS it asks, after hosting, implementation, and delivery or support costs that belong in COGS, what share of each dollar is left to fund growth and eventually profit. Bessemer reports it on company financials. It is a delivery-cost tripwire, not a success score, not Rule of 40, and not LTV:CAC. Write the bar before you measure.
What is a good gross margin for a SaaS / cloud company?
This page does not publish a universal “good” grade table beyond the sources. Bessemer Scaling: average 65–70% regardless of maturity; middle 50% ~60–80%; averages by ARR band 70 / 70 / 65 / 65 / 70; medians 72 / 70 / 65 / 70 / 70; top-performing floors 85%+ / 80% / 75%+ / 80%+ / 80%+. CEO guide: high-growth public-cloud median 77%. Twilio sat ~50% with telecom in COGS — low GM is not automatic failure. Write your own tripwire. Do not invent Cloud 100 LinkedIn medians.
How is gross margin different from Rule of 40, LTV:CAC, and CAC payback?
Rule of 40 is growth % + profit % (or FCF). GM is the delivery-cost ceiling on how high that profit can go. LTV:CAC is lifetime value ÷ CAC; the LTV numerator is often GM-adjusted. CAC payback is months of contribution to recover CAC. Use this page for the GM% sitting; use those pages for those instruments.
How is gross margin different from Magic Number, burn multiple, NRR, and Quick Ratio?
Magic Number is how efficiently prior-period S&M turned into incremental subscription revenue. Burn multiple is company-level Net Burn ÷ Net New ARR. NRR / GRR is cohort retention of a starting ARR base. Quick Ratio is same-period (new + expansion) ÷ (churned + contraction). Gross margin is (Revenue − COGS) ÷ Revenue. They answer different decisions.
Is a Yibud score the same as gross margin? Does it predict success?
No and no. Yibud scores come from a deterministic rule engine on your idea inputs. They point at the weakest assumption to test next. They are not gross margin, not Rule of 40, not LTV:CAC, not CAC payback, not Magic Number, not burn multiple, not NRR, not Quick Ratio, and not a forecast. Neither a GM pass nor a Yibud report score predicts success. You do not need an account to use this page.
Write the gross-margin tripwire before you scale spend
Yibud scores weak assumptions with a deterministic engine. You write signal, threshold, definition, period, and action while you are still honest — then a quarter of Revenue minus COGS is a tripwire, not a 72% slide.
Generate a Yibud report →