◆ Spending benchmarks
What teams actually pay — not what vendors quote.
Pulled from anonymised user stacks, survey data and editorial research. Every table below states the segment it measures, the number of observations behind each row, and where the figure came from.
Benchmark 01 · by team size
Average SaaS spend per employee, monthly
Per-seat SaaS cost roughly triples between micro-teams and the 500+ band. The 11-50 band shows the steepest jump because growing teams add specialist tools (sales engagement, BI, observability) before they hit the volume discount thresholds that larger companies negotiate.
| Team size | Value | Sample |
|---|---|---|
| 1-10 | 214 | n=312 |
| 11-50 | 318 | n=487 |
| 51-200 | 446 | n=341 |
| 201-500 | 512 | n=182 |
| 501+ | 603 | n=94 |
Source · Anonymised user stacks Total sample 1,416
Benchmark 02 · by team size
Number of SaaS tools per company, by team size
Tool count grows roughly 1.4x faster than headcount once a team passes 50 people. Most of the new entries are department-specific niche tools that never get audited until the renewal cycle. Founders running consolidated stacks under 30 tools tend to report higher satisfaction in the same dataset.
| Team size | Value | Sample |
|---|---|---|
| 1-10 | 17 | n=312 |
| 11-50 | 42 | n=487 |
| 51-200 | 89 | n=341 |
| 201-500 | 138 | n=182 |
| 501+ | 211 | n=94 |
Source · Anonymised user stacks Total sample 1,416
Benchmark 04 · by category
Average annual contract value, B2B SaaS
Observability vendors command the highest annual contract values because pricing scales with log volume rather than seats. CRMs sit in the middle of the pack but often add per-feature line items that push effective ACV 30-40% higher by year two.
| Category | Value | Sample |
|---|---|---|
| CRM | 12,400 | n=268 |
| Marketing automation | 9,800 | n=214 |
| BI / analytics | 18,600 | n=143 |
| DevOps / observability | 21,200 | n=127 |
| HR / payroll | 7,600 | n=189 |
| Customer support | 8,400 | n=172 |
Source · Survey data Total sample 1,113
Benchmark 05
Top 5 SaaS spend categories
Sales and marketing together account for 43% of total SaaS spend across the dataset. Engineering tooling under-indexes because cloud infrastructure (AWS, GCP) is excluded — when included it would add another ~22 percentage points and push engineering to the top.
| Segment | Value | Sample |
|---|---|---|
| Sales / CRM | 24 | n=1,416 |
| Marketing & ads tooling | 19 | n=1,416 |
| Productivity & collaboration | 16 | n=1,416 |
| Engineering & infra | 15 | n=1,416 |
| Finance / RevOps | 11 | n=1,416 |
Source · Anonymised user stacks Population 1,416
Benchmark 06 · by tier
Cancellation rate, year 1 of contract
Annual contracts cut year-one churn by roughly two-thirds versus monthly. The trade-off: teams locked into annual deals report 40% more buyer's remorse in qualitative interviews, especially when the tool turns out to be a poor fit by month four.
| Tier | Value | Sample |
|---|---|---|
| Free trial | 62 | n=2,104 |
| Monthly | 34 | n=1,287 |
| Annual | 11 | n=946 |
Source · Editorial research Total sample 4,337
Benchmark 07 · by company size
Discount % obtained when negotiating annual deals
Negotiating leverage scales roughly linearly with seat count. Companies under 10 employees rarely break double-digit discounts, while teams above 500 routinely extract a third off list. The fastest gains come from threatening a competitor switch within 60 days of renewal.
| Company size | Value | Sample |
|---|---|---|
| 1-10 | 8 | n=312 |
| 11-50 | 14 | n=487 |
| 51-200 | 22 | n=341 |
| 201-500 | 31 | n=182 |
| 501+ | 38 | n=94 |
Source · Survey data Total sample 1,416
Benchmark 08 · by team type
Time-to-roll-out new SaaS tool
Sales teams roll out new tools fastest because the buying decision is usually made by the same person who will use them daily. HR and finance lag because rollouts trigger compliance reviews, vendor security questionnaires, and cross-department training that engineering teams typically skip.
| Team type | Value | Sample |
|---|---|---|
| Engineering | 18 | n=214 |
| Marketing | 11 | n=341 |
| Sales | 7 | n=287 |
| Finance | 24 | n=118 |
| HR / People | 28 | n=143 |
Source · Survey data Total sample 1,103
Provenance
Every figure carries its source.
Three kinds of input feed this page, and no table mixes them. The label under each table tells you which one you are reading.
Survey data
Self-reported answers from SaaS buyers. Useful for behaviour and negotiation outcomes that never appear in billing data, with the caveats self-reporting always carries.
3 benchmarksAnonymised user stacks
Aggregated from the anonymised tool stacks buyers have shared with us. Nothing identifies a company, and nothing is attributed to a named vendor.
3 benchmarksEditorial research
Compiled by our editors rather than read straight off a single measured field. The sample column still states the observations behind each row.
2 benchmarksNext
Benchmarks tell you whether. These tell you what.
Price-hike tracker
Which vendors raised prices, by how much, and when — every entry with its source. The other half of a budget conversation.
See the hikesBrowse verified deals
If a benchmark says you are paying over the odds, this is where the discounts are — each one checked before it goes up.
Browse dealsLook up the vocabulary
ACV, contraction MRR, payback period — the terms these tables use, defined in plain English.
Open the glossaryReading these honestly
What the numbers are, what they are not, and how to quote them.
Ask us something elseEach benchmark carries its own source, printed under its table: anonymised user stacks, survey data, or editorial research. Nothing here is a vendor-supplied figure, and no benchmark blends sources — one table, one source.
It is the number of observations behind that row. Where every segment of a benchmark reports the same figure, the same population has been measured several ways, so the total is that figure rather than the sum of the rows.
Probably not exactly. These are central tendencies across a mixed population, not quotes. Treat them as a sanity check on a renewal or a proposal: a number well outside the band for your team size is worth a question, not a panic.
Each benchmark records the date it was last computed, and that date is shown against the table when one exists. Where no date is shown, the figure has not been recomputed since it was first published — so read it as a baseline rather than a live feed.
Yes, with attribution and a link back to this page. Cite the benchmark name and its sample size alongside the number, because a figure quoted without its n is not much of a figure.