Short answer
This report publishes the spread of results people get from the rahuldsarker.co calculators, by metric, once each has 30 anonymised runs. As of 22 October 2026, 0 runs are recorded and no metric has reached the minimum yet, so no figures are shown.
Calculator benchmarks at a glance
Live figures from the dataset, updated 22 October 2026. Each metric shows a median and range only once it has 30 results.
| ROAS | Paid media | 0 | Collecting (0 of 30) | n/a | n/a |
|---|---|---|---|---|---|
| LTV:CAC ratio | Unit economics | 0 | Collecting (0 of 30) | n/a | n/a |
| CAC payback | Unit economics | 0 | Collecting (0 of 30) | n/a | n/a |
| Monthly customer churn | Retention | 0 | Collecting (0 of 30) | n/a | n/a |
| Net revenue retention | Retention | 0 | Collecting (0 of 30) | n/a | n/a |
| SaaS gross margin | SaaS efficiency | 0 | Collecting (0 of 30) | n/a | n/a |
| Rule of 40 score | SaaS efficiency | 0 | Collecting (0 of 30) | n/a | n/a |
| SaaS Magic Number | SaaS efficiency | 0 | Collecting (0 of 30) | n/a | n/a |
1. Why a benchmark built from calculator results?
Most published benchmarks come from surveys or from a vendor’s customer base. This one comes from the numbers people type into free calculators when they are checking their own business.
Survey benchmarks ask people what their numbers are; vendor benchmarks describe companies that already pay for one product. Both are useful, and both are covered in the 2026 marketing benchmarks report with their sources.
Calculator results are a third kind of evidence. They show what founders, marketers and finance teams actually get when they work a number out: the ROAS they see after entering spend and revenue, the LTV : CAC their own churn and margin produce, the payback their CAC implies. Nobody is answering a question about themselves for a survey; they are doing arithmetic for their own decision.
That makes the data candid, but not representative. People reach for a payback calculator when payback worries them, and some runs are what-if scenarios rather than real figures. The report therefore publishes ranges rather than single targets, sets a minimum sample before any figure appears, and keeps the rule-of-thumb bands the calculators already use beside every distribution, so you can see where results fall against a familiar yardstick.
- Seeing the spread of results people get, not one target
- Placing your number in that spread
- Spotting which bands most results fall into
- Tracking how results shift as the dataset grows
- What a typical company in your sector achieves
- Whether any single run was real or a scenario
- Differences by industry or company size (not recorded)
- Anything about who ran the calculator
2. How the data is collected, in brief
Only the graded result is kept: no inputs, no identifiers.
- A result is recorded when a calculator shows one of the twelve benchmarked metrics in this report.
- On the website, only after the visitor has changed at least one input, and only once the numbers have stopped changing for four seconds, so default values and half-typed numbers are not counted.
- Through the MCP server, when an AI assistant runs a benchmarked calculator for its user.
- Each record keeps the day, the calculator, website or MCP, the two-letter country code, and the result rounded to three significant figures with its band.
- No inputs, IP address, visitor or session id, email or company is stored with a result, and nothing links a result to a form submission.
- A metric is published only after 30 results; values outside a plausible range (for example a ROAS above 100x) are dropped as typos.
What each chapter covers
One chapter per metric: definition and formula, the live distribution once it is published, the bands results are graded against, and the calculator to run your own numbers.
Rahul D Sarker, "Calculator Benchmarks Report", rahuldsarker.co, updated 22 October 2026, https://rahuldsarker.co/calculator-benchmarks-report. Quote the sample size (n) with any figure. The published figures are available as a CSV file.
Read the full calculator benchmarks report
The dataset status, the method and the chapter list above are open to everyone. The full report is free: add your details once and every chapter opens in this browser.
- 12 metric chapters with definitions, formulas and the live distribution
- Band shares for every published metric
- The grading bands, by segment where they differ
- A calculator in every chapter to run your own numbers
- The full method, limitations and privacy design
Get the full report
Free. It takes under a minute.
3. ROAS: what calculator users get
Paid media. Figures appear here once 30 anonymised results are in; 0 so far.
What it is. Revenue attributed to advertising for each unit of ad spend.
How it is calculated. ROAS = revenue from ads ÷ ad spend. The ROAS-to-MER converter reports blended MER (total revenue ÷ total marketing spend), graded on the same bands.
Reading it. ROAS only means something next to gross margin. Break-even ROAS is 1 ÷ gross margin, so a 50% margin product needs 2x just to cover the media; when the calculator has your margin, the band is judged against that break-even point rather than the generic bands below.
No figures are shown for ROAS until 30 anonymised runs are in, so a handful of results cannot be mistaken for a benchmark. This section fills in on its own as people use the calculators below; the page refreshes every six hours.
| Losing money | under 1x | Ads return less revenue than they cost. |
|---|---|---|
| Likely unprofitable | 1x to 2x | Unprofitable at most product margins. |
| Typical | 2x to 4x | Workable; profit depends on margin. |
| Strong | 4x and above | Room to scale spend. |
- ROAS to MER Converter also reports ROAS.
4. LTV:CAC ratio: what calculator users get
Unit economics. Figures appear here once 30 anonymised results are in; 0 so far.
What it is. How many times a customer’s lifetime gross profit covers the cost of acquiring them.
How it is calculated. LTV : CAC = (average revenue per customer × gross margin × expected lifetime) ÷ customer acquisition cost.
Reading it. The ratio is only as good as the lifetime estimate behind it, and lifetime is usually derived from churn. A young business with a few months of data can show an attractive ratio that rests on an assumed lifetime nobody has observed yet.
No figures are shown for LTV:CAC ratio until 30 anonymised runs are in, so a handful of results cannot be mistaken for a benchmark. This section fills in on its own as people use the calculators below; the page refreshes every six hours.
| Losing money | under 1x | Each customer costs more to win than they will ever return. |
|---|---|---|
| Below target | 1x to 3x | Growth is buying customers at thin or negative returns; fix retention or CAC before scaling spend. |
| Healthy | 3x to 5x | Customers pay back acquisition cost with room to reinvest. |
| Very efficient | 5x and above | Unusually high: you are likely under-investing in growth and could spend more to win share. |
5. CAC payback: what calculator users get
Unit economics. Figures appear here once 30 anonymised results are in; 0 so far.
What it is. Months of gross profit from a new customer needed to earn back what it cost to acquire them.
How it is calculated. CAC payback = CAC ÷ (monthly revenue per customer × gross margin).
Reading it. Payback measures cash, not value: a long payback can still be a good investment if customers stay, but it ties up cash that has to be financed. The acceptable range depends on deal size, so the bands differ for SMB, mid-market and enterprise.
No figures are shown for CAC payback until 30 anonymised runs are in, so a handful of results cannot be mistaken for a benchmark. This section fills in on its own as people use the calculators below; the page refreshes every six hours.
| SMB | Excellent | under 6 months | Fast payback frees cash to reinvest. |
|---|---|---|---|
| SMB | Healthy | 6 months to 12 months | Within the 12-month SMB norm. |
| SMB | Slow | 12 months to 18 months | Cash is tied up too long for SMB churn levels. |
| SMB | Too slow | 18 months and above | Customers may churn before they pay back. |
| Mid-market | Excellent | under 9 months | Fast payback for mid-market. |
| Mid-market | Healthy | 9 months to 18 months | Within the 18-month mid-market norm. |
| Mid-market | Slow | 18 months to 24 months | Growth will need outside cash. |
| Mid-market | Too slow | 24 months and above | Acquisition is burning cash faster than it returns. |
- LTV : CAC Ratio Calculator also reports CAC payback.
6. Monthly customer churn: what calculator users get
Retention. Figures appear here once 30 anonymised results are in; 0 so far.
What it is. The share of customers lost each month.
How it is calculated. Monthly churn = customers lost in the month ÷ customers at the start of the month.
Reading it. Small monthly numbers compound: 3% a month loses roughly 31% of customers over a year. Logo churn and revenue churn can tell different stories when large and small customers behave differently.
No figures are shown for Monthly customer churn until 30 anonymised runs are in, so a handful of results cannot be mistaken for a benchmark. This section fills in on its own as people use the calculators below; the page refreshes every six hours.
| SMB | Excellent | under 1% | Rare for SMB. |
|---|---|---|---|
| SMB | Healthy | 1% to 3% | At or better than typical SMB churn. |
| SMB | High | 3% to 5% | Typical-to-high for SMB; retention is capping LTV. |
| SMB | Severe | 5% and above | More than half the base turns over each year. |
| Mid-market | Excellent | under 0.7% | Strong for mid-market. |
| Mid-market | Healthy | 0.7% to 1.5% | In line with mid-market norms. |
| Mid-market | High | 1.5% to 2.5% | Above mid-market norms. |
| Mid-market | Severe | 2.5% and above | A retention problem, not a growth problem. |
- LTV : CAC Ratio Calculator also reports Monthly customer churn.
7. Net revenue retention: what calculator users get
Retention. Figures appear here once 30 anonymised results are in; 0 so far.
What it is. Revenue kept from an existing cohort after a year, including expansion, contraction and churn.
How it is calculated. NRR = (starting recurring revenue + expansion − contraction − churned revenue) ÷ starting recurring revenue.
Reading it. Above 100% the existing base grows on its own; below it, new sales first have to replace losses. Like payback, the bands differ by segment because larger customers have more room to expand.
No figures are shown for Net revenue retention until 30 anonymised runs are in, so a handful of results cannot be mistaken for a benchmark. This section fills in on its own as people use the calculators below; the page refreshes every six hours.
| SMB | Poor | under 80% | The base shrinks fast; new sales mostly replace losses. |
|---|---|---|---|
| SMB | Below median | 80% to 95% | Below typical SMB retention. |
| SMB | Healthy | 95% to 105% | Solid for SMB. |
| SMB | Best in class | 105% and above | The base grows on its own. |
| Mid-market | Poor | under 90% | The base is shrinking. |
| Mid-market | Below median | 90% to 100% | Expansion is not covering churn. |
| Mid-market | Healthy | 100% to 115% | Expansion outpaces churn. |
| Mid-market | Best in class | 115% and above | Top-quartile expansion. |
8. SaaS gross margin: what calculator users get
SaaS efficiency. Figures appear here once 30 anonymised results are in; 0 so far.
What it is. Revenue left after the direct cost of delivering the product.
How it is calculated. Gross margin = (revenue − cost of revenue) ÷ revenue. Cost of revenue usually includes hosting, third-party software in the product, support and onboarding.
Reading it. Definitions vary most here: whether customer success, onboarding or payment fees sit in cost of revenue can move margin by several points, so compare like with like.
No figures are shown for SaaS gross margin until 30 anonymised runs are in, so a handful of results cannot be mistaken for a benchmark. This section fills in on its own as people use the calculators below; the page refreshes every six hours.
| Low for SaaS | under 60% | Closer to services economics; valuation multiples suffer. |
|---|---|---|
| Below median | 60% to 70% | Hosting or support costs are heavy. |
| Healthy | 70% to 80% | Around the SaaS median. |
| Best in class | 80% and above | Top-quartile software margins. |
9. Rule of 40 score: what calculator users get
SaaS efficiency. Figures appear here once 30 anonymised results are in; 0 so far.
What it is. Growth rate plus profit margin, as one score.
How it is calculated. Rule of 40 score = annual revenue growth % + profit margin % (often EBITDA or free cash flow margin).
Reading it. The score lets a fast-growing, loss-making company and a slow, profitable one be compared on one axis. It says nothing about which of the two is the better path for a given business.
No figures are shown for Rule of 40 score until 30 anonymised runs are in, so a handful of results cannot be mistaken for a benchmark. This section fills in on its own as people use the calculators below; the page refreshes every six hours.
| Well below | under 20% | Neither growing fast nor profitable. |
|---|---|---|
| Below 40 | 20% to 40% | Growth plus margin falls short of the investor bar. |
| Passing | 40% to 60% | Meets the bar investors use for healthy SaaS. |
| Elite | 60% and above | Top-tier balance of growth and profit. |
10. SaaS Magic Number: what calculator users get
SaaS efficiency. Figures appear here once 30 anonymised results are in; 0 so far.
What it is. New annualised revenue produced by each unit of sales and marketing spend.
How it is calculated. Magic Number = (this quarter’s recurring revenue − last quarter’s) × 4 ÷ last quarter’s sales and marketing spend.
Reading it. Quarter-to-quarter noise is large for small companies; one big deal can swing the number, so read it over several quarters.
No figures are shown for SaaS Magic Number until 30 anonymised runs are in, so a handful of results cannot be mistaken for a benchmark. This section fills in on its own as people use the calculators below; the page refreshes every six hours.
| Inefficient | under 0.5x | Fix conversion or retention before adding spend. |
|---|---|---|
| Acceptable | 0.5x to 0.75x | Investigate before scaling spend. |
| Efficient | 0.75x to 1.5x | Safe to invest more in sales and marketing. |
| Very efficient | 1.5x and above | Under-spending: there is room to press harder. |
11. SaaS quick ratio: what calculator users get
SaaS efficiency. Figures appear here once 30 anonymised results are in; 0 so far.
What it is. Recurring revenue added for every unit lost.
How it is calculated. Quick ratio = (new MRR + expansion MRR) ÷ (churned MRR + contraction MRR).
Reading it. A high ratio with little absolute growth is possible when losses are tiny; read it with the growth rate, not instead of it.
No figures are shown for SaaS quick ratio until 30 anonymised runs are in, so a handful of results cannot be mistaken for a benchmark. This section fills in on its own as people use the calculators below; the page refreshes every six hours.
| Shrinking | under 1x | Losing more MRR than you add. |
|---|---|---|
| Weak | 1x to 2x | Growth is leaky. |
| Healthy | 2x to 4x | Adding clearly more than you lose. |
| Excellent | 4x and above | Efficient, compounding growth. |
12. Website conversion rate: what calculator users get
Website and funnel. Figures appear here once 30 anonymised results are in; 0 so far.
What it is. The share of visitors who complete the goal: a purchase, a lead or a sign-up.
How it is calculated. Conversion rate = conversions ÷ visitors (or sessions) in the same period.
Reading it. Traffic mix moves this more than page design: branded and returning visitors convert far better than cold paid traffic, so a site-wide rate hides more than it shows.
No figures are shown for Website conversion rate until 30 anonymised runs are in, so a handful of results cannot be mistaken for a benchmark. This section fills in on its own as people use the calculators below; the page refreshes every six hours.
| Low | under 1% | Traffic, offer or page is misaligned. |
|---|---|---|
| Typical | 1% to 2.5% | Around the cross-industry median. |
| Strong | 2.5% to 5% | Well above median. |
| Top decile | 5% and above | Exceptional; check the traffic is not mostly branded. |
13. Marketing budget as % of revenue: what calculator users get
Planning. Figures appear here once 30 anonymised results are in; 0 so far.
What it is. Marketing spend as a share of revenue.
How it is calculated. Marketing budget % = total marketing spend ÷ revenue for the same period.
Reading it. The right level depends on growth targets, gross margin and payback, not on what other companies spend; the bands describe common ranges, not targets.
No figures are shown for Marketing budget as % of revenue until 30 anonymised runs are in, so a handful of results cannot be mistaken for a benchmark. This section fills in on its own as people use the calculators below; the page refreshes every six hours.
| Lean | under 5% | Maintenance-level spend; hard to take share. |
|---|---|---|
| Typical | 5% to 12% | In the common range for growing companies. |
| Aggressive | 12% to 20% | Growth-stage spend; needs tight unit economics. |
| Very aggressive | 20% and above | Only sustainable with strong payback. |
14. Net Promoter Score: what calculator users get
Customer. Figures appear here once 30 anonymised results are in; 0 so far.
What it is. How likely customers are to recommend the company, on a scale from −100 to 100.
How it is calculated. NPS = % promoters (scores 9–10) − % detractors (scores 0–6).
Reading it. Response rates and who is asked shape NPS as much as loyalty does. Trends within one company are more useful than comparisons between companies.
No figures are shown for Net Promoter Score until 30 anonymised runs are in, so a handful of results cannot be mistaken for a benchmark. This section fills in on its own as people use the calculators below; the page refreshes every six hours.
| Poor | under 0 | More detractors than promoters. |
|---|---|---|
| Good | 0 to 30 | Positive, room to improve. |
| Great | 30 to 50 | Strong word of mouth. |
| Excellent | 50 to 70 | Exceptional loyalty. |
| World class | 70 and above | Rare territory. |
15. Which calculators produce the most results?
Recorded results by calculator, website and MCP combined. Only calculators with a benchmarked metric are counted.
This table lists the calculators that have produced the most results once recording begins.
16. Method and limitations in full
What is counted, what is dropped, and the biases to keep in mind before quoting any figure.
Grading. Each result is graded on the server against the same bands the calculators show, recomputed from the value rather than taken from the browser. Where a metric has different bands by segment (CAC payback, churn and NRR), the segment the visitor selected is used, SMB by default. For ROAS, when the calculator has the gross margin, the band is judged against break-even ROAS for that margin.
Percentiles. Medians and the 25th and 75th percentiles are calculated across every recorded result for the metric, with linear interpolation, and rounded for display. Band shares are the percentage of results in each band.
Minimum sample. Nothing is shown for a metric with fewer than 30 results. At small samples the percentiles still move noticeably with each new result, so quote the n alongside any figure.
Refresh. Figures are recalculated at most every six hours; the date at the top of the page is the last refresh.
- Self-selection: people tend to open a calculator when a number worries them, which can skew results towards weaker bands.
- Scenarios: some runs are what-ifs, not actual figures. The four-second settle rule removes half-typed values, not hypotheticals.
- Repeat runs: one person trying several scenarios adds several results. No identifier is kept, so these cannot be removed without tracking people.
- AI assistant runs: results through the MCP server use whatever numbers the assistant was given, which may be examples rather than a real business.
- No segmentation by industry, size or market is possible, because none of that is recorded.
17. What is stored, and what is not
The dataset was designed so that no row can be tied back to a person or a company.
- Day (no time)
- Calculator
- Website or MCP
- Two-letter country code
- Benchmarked result, rounded to 3 significant figures
- Its band and segment
- The inputs you typed
- IP address, visitor or session id
- Name, email, phone or company
- Any link to forms, bookings or reports you request
Visits by the site owner while signed in to the dashboard are excluded. Recording can be switched off for the whole site at any time. See the privacy policy for how the rest of the site handles data.
Want a second opinion on your own numbers?
The distribution tells you where a number sits. A review looks at why, and what to change first.
Bring your calculator results; the call starts from your numbers.
Calculator benchmarks: common questions
Where does the data in this report come from?
From results people get on the rahuldsarker.co calculators and from AI assistants running the same calculators through its MCP server. Only the graded result is stored: the day, the calculator, website or MCP, the country code and the rounded result with its band.
Are my calculator inputs stored?
No. Only the benchmarked result, such as a ROAS of 3.2x, is kept, rounded to three significant figures, with no inputs, IP address, visitor id or contact details.
Why are some metrics not shown yet?
A metric is published only once it has 30 results, so a few runs cannot pass for a benchmark. Sections fill in automatically as results come in.
Is this a representative benchmark for my industry?
No. Calculator users choose to run the numbers, some runs are scenarios, and no industry or company size is recorded. Use the ranges to place a number, and the sourced reports on this site for sector benchmarks.
How often is the report updated?
Figures are recalculated at most every six hours, and the CSV file updates with them.
What are the bands each result is graded against?
The rule-of-thumb ranges the calculators show, such as an LTV : CAC of 3x to 5x as healthy. Each chapter lists the bands and their basis; they frame a conversation and are not a dataset.
Can I download the data?
Yes. The published figures are in a CSV file at https://rahuldsarker.co/calculator-benchmarks-report/key-figures.csv. Cite the report with the sample size.
Sources
- Anonymised calculator results, rahuldsarker.co (live dataset, CSV), RDS & CO. Accessed 2026-10-22.