Analytics

Marketing ROI calculators: the maths your ad platform hides

Four working calculators, and the reason the number on your dashboard is almost always the flattering one.

Editorial line illustration: a two-axis chart with two curves diverging, one for reported return and one for true return, linked by a wavy line to a calculator with a single orange key, beside a tipping balance scale, on cream paper.

▸ Bottom line up front

The return on ad spend your platform reports is attributed revenue, not incremental revenue, and the gap is rarely small. Stella's 2025 incrementality benchmarks, built from 225 geo tests run August 2024 to December 2025, found platform-reported ROAS overstates true incremental ROAS by 2 to 3 times on average, and 5 to 10 times on branded search and retargeting. The four calculators below run the honest version of the maths: incremental ROAS, CAC payback on gross margin, gross-margin LTV:CAC, and a lead-to-close unit-economics chain. Use them, then read why the platform's own AI calculator quietly rounds up.

Reported ROAS vs the ROAS you actually earned

Platform-reported return on ad spend (ROAS) is attributed revenue divided by ad spend, and the platform picks which revenue counts. True incremental ROAS (iROAS) is what you would lose if you switched the channel off, measured by holding a group out and comparing. The first number is a marketing report. The second is the one your finance team cares about.

The honest formula is iROAS = (test revenue - control revenue) / test spend. You cannot get it from a dashboard, only from a geo or audience holdout. The calculator below lets you take a platform-reported figure and discount it by an incrementality factor so you can see the phantom revenue you have been counting. Edit every field. The defaults are an example, not a leapbuzz benchmark.

▸ Calculator 1 · Reported ROAS vs incremental ROAS

0.60
Reported ROAS4.00x
Incremental ROAS2.40x
Phantom revenue$16,000

I ask one question before any budget meeting: which number in this deck would survive a holdout test? Platform ROAS flatters whoever spent the most. Incremental ROAS tells you which dollar bought a customer you would not have got anyway. The first number wins the meeting. The second one grows the business. Build the calculator, and the budget, around the second.
Siddharth Surana
Founder, leapbuzz
18+ years in marketing and digital leadership

CAC payback, and the gross-margin trap

Customer acquisition cost (CAC) payback period is the number of months to earn back what you spent acquiring a customer. The trap is running it on revenue. You recover CAC out of gross profit, not top-line revenue, so the correct formula is CAC / (monthly ARPA × gross margin) where ARPA is average revenue per account.

The benchmark everyone quotes, under 12 months, is now the top of the class. The 2024 SaaS Benchmarks from High Alpha, run with OpenView and Paddle across more than 800 companies, put the median CAC payback around 18 months, up from 14 the year before. KeyBanc's 2024 survey put its median near 20 months. So a 14-month payback is healthy today, not alarming.

▸ Calculator 2 · CAC payback period (gross-margin aware)

75%
Gross profit / month$112.50
Payback (months)10.7
vs revenue-only8.0

CAC payback period read bands (2026)
PaybackReadWhat to do
Under 12 monthsBest-in-classConsider spending faster while the maths holds.
12 to 18 monthsHealthy, at the 2024 medianHold the line, watch margin and churn.
18 to 24 monthsWorth watchingAudit channel mix and gross margin before scaling.
Over 24 monthsCash-flow riskFix unit economics before adding budget.

LTV:CAC done on profit, not revenue

The 3:1 lifetime-value-to-CAC rule comes from David Skok's SaaS metrics writing on the For Entrepreneurs blog: below 3:1 your unit economics are too weak to fund growth, well above it you may be under-investing. The catch is that the L in LTV must be gross profit. Gross-margin LTV is (ARPA × gross margin) / monthly churn. Run it on revenue and you inflate the ratio by one over your margin, so a 50 percent-margin business bragging about 3:1 is really sitting at 1.5:1.

KeyBanc's 2024 SaaS survey confirms companies overwhelmingly compute LTV on gross profit for exactly this reason. The calculator shows both numbers side by side so the inflation is visible. Skok himself warns the rule was never meant for pre-product-market-fit companies, and it says nothing about cash timing, so read it next to your payback period, never alone.

▸ Calculator 3 · LTV:CAC (gross-margin adjusted)

3.0%
75%
Gross-margin LTV$3,750
LTV:CAC (honest)3.1:1
LTV:CAC on revenue4.2:1

The lead-to-close chain a CFO will believe

For pipeline businesses, ROI lives in the funnel, not the click. Closed-won deals are leads × (lead to MQL) × (MQL to SQL) × (SQL to close), where MQL is a marketing-qualified lead and SQL is a sales-qualified lead. Multiply the stage rates, divide your paid spend by the deals that result, and you get paid CAC, the number that matters for budget decisions.

Do not confuse paid CAC with blended CAC. Blended CAC mixes in organic, referral and word-of-mouth customers who cost no media, so it is always lower and it always flatters paid channels. Quote blended CAC to justify ad budget and you will over-invest in your weakest channel. This calculator runs the paid chain on gross profit so the ROI multiple is the one finance will sign off.

▸ Calculator 4 · Lead-to-close unit economics

40%
50%
25%
Closed-won deals20
Paid CAC$1,000
ROI multiple3.0x

What platform offers AI-driven ROI calculators for ad campaigns?

No major ad platform ships a literal profit ROI calculator, because platforms measure revenue, not profit. What they ship is AI-driven forecasting of conversions and conversion value, and several of them are genuinely useful inside their own walls. The catch is universal: every one of them forecasts against the conversions you already track, so the output inherits your attribution model and never subtracts the baseline of what would have happened anyway.

Where platform ROI forecasters help and where they mislead
ToolWhere it genuinely helpsWhere it misleads
Google Ads Performance PlannerForecasts clicks, conversions, conversion value and ROAS from your account history and simulated auctions, refreshed about every 24 hours. Good for spotting diminishing returns inside Google.Forecasts against your tracked conversions, so the projection carries your attribution, and it cannot see cannibalised organic or cross-platform overlap.
Meta Ads Manager estimatesEstimated results during setup give a relative read on reach and likely cost per result inside Meta.Estimates reflect Meta's default 7-day click plus 1-day view attribution, which runs hot on retargeting and low-funnel campaigns.
Microsoft Advertising plannerMirrors Google's planner for the Microsoft and LinkedIn-fed search network, useful for B2B budget planning.Same attribution dependence as Google; not an incrementality measure.
HubSpot and CRM analyticsTies ad spend to closed-won deals, which is closer to real pipeline ROI than front-end ROAS.Only as honest as the attribution and stage definitions you feed it; still no holdout baseline.
Lift and experiment toolsMeta Conversion Lift and geo experiments are the only native features that measure true incrementality.Need real volume and correct randomisation, and most advertisers skip the design work.

So the honest answer to the query is: use Performance Planner and its peers for intra-platform budget allocation, where they are strong, and never read their output as your business ROI. For the money question, the only tools that count are incrementality experiments and marketing mix modelling. If you want the AI to do real ROI work rather than flattering forecasts, that is where an AI marketing strategy engagement earns its keep, and where the right Google and Meta setup, plus a credible search advertising structure, stops the platforms from grading their own homework.

How to sanity-check any ROI calculator

Whether it is a platform tool, a spreadsheet, or one of the four above, three questions tell you whether to trust a calculator's output as ROI. Ask them before you move budget.

  1. What attribution feeds it? If the answer is last-click or platform-attributed, the number is descriptive, not causal. It tells you what got credit, not what caused the sale.
  2. Does it use gross margin or revenue? Revenue-based ROI, LTV and payback are inflated by one over your margin. A calculator that never asks for your margin is quietly optimistic.
  3. Does it subtract a baseline? An honest tool nets out the conversions that would have happened anyway. View-through conversions and branded search are where this matters most, and where most tools simply count everything.

One more structural point. The same conversion often gets claimed twice. Meta counts a 7-day click plus 1-day view, Google leans on last-click, and the two do not share an identity graph, so a buyer who clicks a Meta ad on Monday and searches your brand on Wednesday is counted by both. Sum the dashboards and your reported conversions exceed reality. Deduplicate against one source of truth, then validate channel lift with holdouts. Getting that source of truth right, and the website and tracking under it, is often a build job rather than a media job; we treat website development as part of the same engagement and you can talk to us about it. The proof that this discipline pays out is on our results page, and the reporting cadence that keeps it honest is in GA4 actionable insights.

Questions, answered.

What platform offers AI-driven ROI calculators for ad campaigns?

No major ad platform ships a literal profit ROI calculator, because platforms measure revenue (ROAS), not profit. What they ship is AI-driven forecasting of conversions and conversion value. Google Ads Performance Planner is the canonical one: per Google Ads Help, it forecasts clicks, conversions, conversion value, and impressions using your account history and simulated auctions, refreshed roughly every 24 hours. Meta Ads Manager shows estimated results during campaign setup, Microsoft Advertising mirrors Google's planner, and CRM tools like HubSpot tie ad spend to closed-won deals. All of them forecast against the conversions you already track, so the output inherits your attribution model. None subtract the baseline of what would have happened anyway, which is why their ROAS reads optimistic. Treat them as intra-platform budget planners, not measures of true business ROI.

What is the difference between ROAS and incremental ROAS?

ROAS is attributed revenue divided by ad spend, where the platform decides which revenue to attribute. Incremental ROAS (iROAS) is measured by experiment: revenue in a test group minus revenue in a held-out control group, divided by test spend. Only the holdout version is causal. Stella's 2025 DTC incrementality benchmarks, drawn from 225 geo tests run August 2024 to December 2025, found the gap between platform-reported ROAS and true incremental ROAS commonly reaches 2-3x, and 5-10x on branded search and retargeting. Their median incremental ROAS for Meta was 2.92x, while Google branded search came in at 0.70x incremental, far below what the platform dashboard would suggest.

Why must LTV and CAC payback use gross margin instead of revenue?

Because you recover acquisition cost out of gross profit, not revenue. CAC payback period is CAC divided by (monthly ARPA times gross margin), and gross-margin LTV is (ARPA times gross margin) divided by monthly churn. If you run those numbers on revenue alone, you overstate your capacity to repay CAC by a factor of one over your margin. A business with 50 percent margin reporting a 3:1 LTV:CAC on revenue is really at 1.5:1 on gross profit, below the survival line. KeyBanc's 2024 SaaS survey confirms that companies overwhelmingly compute LTV on gross profit for this reason.

What is a good LTV:CAC ratio?

The 3:1 rule comes from David Skok's SaaS Metrics work on the For Entrepreneurs blog: below 3:1 suggests weak unit economics, well above it can mean you are under-investing in growth. The ratio must be calculated on gross-margin LTV, not revenue. Skok himself cautions the rule was never meant for pre-product-market-fit companies, and it says nothing about cash timing: you can hit 3:1 and still have a 30-month payback that starves your cash. Read the ratio alongside CAC payback, never in isolation.

What is a healthy CAC payback period?

The traditional benchmark is under 12 months, but that is now top-quartile or product-led territory. The 2024 SaaS Benchmarks (High Alpha with OpenView and Paddle, 800-plus companies) put the median CAC payback around 18 months, up from 14 the prior year, and KeyBanc's 2024 survey put the median near 20 months. Use under 12 as best-in-class, 12 to 18 as acceptable, 18 to 24 as worth watching, and over 24 as a cash-flow risk. The number is also highly sensitive to gross margin, so always compute it on contribution margin.

Why do Meta and Google sometimes both claim the same conversion?

Because they use different attribution and do not share an identity graph. Meta defaults to a 7-day click plus 1-day view window, Google leans on last-click, and neither knows about the other. A buyer who clicks a Meta ad on Monday and searches your brand on Wednesday gets counted by both. Sum the two dashboards and your total reported conversions exceed the conversions you actually had. This double-counting is the single most common reason a portfolio of campaigns looks profitable on platform numbers and flat on the bank statement. Deduplicate with a single source of truth, then validate channel lift with holdout tests or marketing mix modelling.

How do I sanity-check any marketing ROI calculator?

Ask three questions. First, what attribution feeds it: if the answer is last-click or platform-attributed, the output is descriptive, not causal. Second, does it use gross margin or revenue: revenue-based ROI and LTV are inflated by one over your margin. Third, does it subtract a baseline: an honest calculator nets out the conversions that would have happened anyway. If a tool fails any of the three, keep the directional read for intra-platform decisions and confirm the money question with an incrementality experiment.

Related reading