The Customer Proof Index 2026
What 1,000 B2B software companies actually publish, measured by our own crawler instead of taken on their word.
We crawled 1,000 B2B software companies across 55 categories on 2026-08-21 and found that 39.0% (372 of 954) have no written proof of a customer result anywhere on their site. That splits into 21.5% with nothing at all and 17.5% that substitute video only.
Of the 335 companies where we could open a real story, 1 in 10 (10.7%) contains no number anywhere, even though 84.8% name the customer and 83.6% quote them directly. Most published B2B case studies are personal, just not measurable. And they go stale: among stories with a genuine, machine-readable publish date, the median newest one is about 9.9 months old, and 44.2% of those companies have not published a new story in over a year.
Key findings
- 21.5% of B2B software companies (205 of 954) have no written case study anywhere on their site; another 17.5% (167) substitute video-only proof with no written companion.
- Among the 335 companies with at least one readable story, 10.7% (36) never state a number in any sampled story: no percentage, dollar figure, or before-and-after.
- 84.8% name the customer and 83.6% include a direct quote: the parts of a case study that take the least verification effort are the parts most companies get right.
- The median B2B case study runs 858 words (n = 1,348 stories sampled across the 335 companies).
- IT, dev & security companies are the least likely to have written proof (48.4% have nothing written, n = 221); HR & people companies are the most likely (30.8% have nothing written, n = 91).
- Case studies age fast and are rarely dated: only 120 of 335 story-bearing companies carry a genuine publish date at all, and among those, the median newest story is about 9.9 months old.
How much customer proof do B2B software companies actually publish?
Three states, measured across every company we could reach (n = 954, excluding 46 unreachable homepages). “Written proof” means at least one page reads as a real case study or customer story, not a testimonial wall of one-line blurbs.
It gets thinner once you open the stories
Having a case study page and having a credible one are different questions. Among the 335 companies where at least one sampled story loaded successfully:
10.7%
have no number in any sampled story (36 of 335)
84.8%
name the customer by strict signals (284 of 335)
83.6%
include a direct customer quote (280 of 335)
858
median words per story (n = 1,348 stories)
Which B2B software categories have the least written proof?
We grouped the 55 SaaSHub categories into 7 broader groups and measured “nothing written” (no written case study and no video either) and, within the companies that do have a readable story, the share with no number in it. Sorted worst to best by nothing-written rate.
| Category group | Companies (n) | Nothing written | Of story-bearing, no number |
|---|---|---|---|
| IT, dev & security | 221 | 107 (48.4%) | 10/62 (16.1%) |
| Operations & vertical | 87 | 36 (41.4%) | 1/30 (3.3%) |
| Finance & admin | 212 | 83 (39.2%) | 6/62 (9.7%) |
| Customer support & success | 93 | 35 (37.6%) | 3/40 (7.5%) |
| Marketing | 179 | 61 (34.1%) | 6/77 (7.8%) |
| Sales & CRM | 71 | 22 (31.0%) | 5/28 (17.9%) |
| HR & people | 91 | 28 (30.8%) | 5/36 (13.9%) |
“Nothing written” % is of all companies in the group (n excludes crawl errors). “No number” % is of only the story-bearing subset in that group. See the methodology page for exactly which 55 categories map to which group.
How fresh is the case study a buyer actually finds?
Among the 335 companies with a readable story, we only trust a date if it comes from a genuine, machine-readable source (JSON-LD datePublished, the article:published_time meta tag, or a <time> element). Server timestamps and sitemap timestamps track redeploys, not edits, so we exclude them from this chart even though they would make the numbers look fresher.
Among the 120 companies where we do have a genuine date, the median newest story is about 9.9 months old, and 44.2% have not published a new one in over a year. Read that alongside the other 215: a missing date is not proof of a stale story, only proof we could not verify it either way. Full source breakdown on the methodology page.
What the best do
We did not hand-pick these from memory. Every company below cleared the same objective bar inside our own dataset: a named customer, a real quantified result, a direct quote, and a story with a genuine publish date within the last 12 months. Quotes are shown as our crawler captured them: some are snippet windows around the matched text, and an ellipsis marks where a sentence continues beyond what we captured.
CI/CD tool
bitrise.io →The story about Shopify states that up to 50% faster CI/CD builds and 50% fewer tools.
- Named: named through a JSON-LD Organization record, not just the headline
- Quoted: “We started questioning whether it was better to look outside for an off-the-shelf solution that could offer us more functionality and help improve our merchant experience.”
Published 2026-08-14, about a week before our crawl.
Recruiting tool
recruitee.com →The story about Keolis states that Keolis brought its time-to-hire down to 18 days.
- Named: introduced in an “About Keolis” section
- Quoted: “We were flying blind at the time. We had no compass to tell us where we were heading with recruitment.”
Published 2026-08-21, the same day as our crawl.
Contract management tool
gatekeeperhq.com →The story about Redwood Logistics states that the customer uses the vendor’s agent to return around 300 hours to procurement by automating renewal preparation.
- Named: the customer’s name appears 8 times through the body, not just in the title
- Quoted: “Gatekeeper’s Agent handles the work my team would normally do at the start of every renewal… By the time we start, the groundwork is already done.”
Published 2026-07-18, about five weeks before our crawl.
Fleet management tool
verizonconnect.com →The story about Green Mountain states that puts the number in its headline without hedging: “Green Mountain saves $70k in auto claims”.
- Named: the customer’s name appears 23 times through the piece
- Quoted: “Day to day, there’s a lot of driving…” (the weakest quote of the five; our crawler caught a fragment, not a full sentence)
Published 2026-08-21, the same day as our crawl.
Help desk tool
front.com →The story about Gattinoni Business Travel states that about 65% of Gattinoni’s tickets are classified with enough confidence to route automatically.
- Named: the customer’s name appears 13 times through the piece
- Quoted: “At Gattinoni, great service means combining human expertise with technology…”
Published 2026-06-24, about two months before our crawl.
How we measured this
We are not an independent research firm. StoryVoice sells case-study software, and this data comes from a crawler we built for this report. Four things matter most when deciding how much weight to put on any number above:
- We can only see raw HTML. A proof page that renders its content client-side (common on React or Vue marketing sites backed by a headless CMS) looks empty to this crawler even when a human visitor would see stories. This likely inflates the “no written proof” bucket by an amount we cannot measure.
- The checks are regex and DOM pattern-matching, not a human reading every page. They will miss unusually phrased results and can be fooled by an unrelated capitalized name or logo. We report a stricter and a looser version of the “names the customer” check for this reason; every figure in this report uses the stricter one.
- Publish dates are detected, not guaranteed. Only 120 of the 335 story-bearing companies carry a genuine, machine-readable date. We exclude server and sitemap timestamps from every freshness claim because they track redeploys, not edits, and including them makes the numbers look artificially fresh.
- The sample is one directory's curation, not a census. All 1,000 companies came from 55 SaaSHub “best of” category pages. That is a reasonable cross-section of B2B software that maintains a public listing, but it is not a random sample of every B2B software company, and this report never claims otherwise.
The full method, every heuristic in plain English, the complete date-source table, and per-figure sample sizes are on the methodology page.
What this means for you
- 01You have zero written case studies. You are one of 205 companies (21.5%) in exactly this position, not an outlier. Start with one, in the segment you are most starved for proof in.
- 02You only have video testimonials. 17.5% of companies are in the same spot. Video does not get indexed, quoted, or pasted into a sales email the way a written story does. A written companion to your strongest video is worth more than a second video.
- 03You have case studies, but no numbers in them. You are not alone (10.7% of story-bearing companies), but you are leaving the highest-trust part of a case study out. A before-and-after in the customer's own words is usually sitting in the interview you already did.
- 04You sell IT, dev, security, finance, or operations software. Your category is behind its peers (39-48% have nothing written). That is a bigger opportunity than it is an excuse.
- 05Your newest case study is over a year old. 44.2% of dated stories are in the same position. A stale library reads as an inactive company to a buyer doing due diligence.
Frequently asked questions
How common are case studies in B2B software?
Less common than most marketers assume. In our crawl of 954 B2B software companies (2026-08-21), 61.0% publish at least one written case study, which means 39.0% have nothing written: 21.5% have no customer proof of any kind and 17.5% substitute video testimonials with no written companion. A written case study is the majority position, not a given.
What share of published case studies have a number in them?
About 9 in 10. Among the 335 companies where we could open at least one real story, 89.3% include a quantified result somewhere (a percentage, a dollar figure, a time saved, a before-and-after). The remaining 10.7% (36 companies) never state a number in any story we sampled. Naming and quoting are stronger still: 84.8% name the customer and 83.6% include a direct quote.
How old are the case studies most companies publish?
Older than they look, on the ones we could date at all. Of the 335 story-bearing companies, only 120 carry a genuine, machine-readable publish date (JSON-LD, an article meta tag, or a time element). Among those 120, the median newest story is about 9.9 months old, and 44.2% have not published a new one in over a year. The other 215 companies have no genuine date on their story, so their real freshness is unknown rather than confirmed stale.
Which B2B software categories have the least written proof?
IT, dev and security tools, by a wide margin. 48.4% of companies in that group (107 of 221) have nothing written, the worst of the 7 groups we measured. Operations and vertical software (fleet, property, field service, legal practice management) was second-worst at 41.4%. HR and people software was the strongest group, at 30.8% with nothing written.
How was this data collected?
We crawled the homepages of 1,000 B2B software companies drawn from 55 SaaSHub category listings on 2026-08-21, read each site's navigation and footer for a proof page, and sampled up to 5 individual story pages per company where one existed. Every claim in this report comes from a regex and DOM heuristic applied to raw HTML, not a human reading every page, so it is a directional measurement, not an audit of any single company. The full method, every heuristic, and per-figure sample sizes are on the methodology page.
Cite this research
StoryVoice Research. (2026, August 21). The Customer Proof Index 2026: What 1,000 B2B Software Companies Actually Publish. StoryVoice. https://storyvoice.io/research/customer-proof-index-2026
This is our own crawl, not a syndicated survey. If you are citing a specific figure, please link to this page or the methodology rather than restating the number without its source. Every figure here carries its own n, and pulling one out of context (for example, quoting “61% have proof” without the 39% who do not) changes what it means.
Data in this report comes from a first-party crawl of 1,000 B2B software company homepages, conducted by StoryVoice on 2026-08-21, sourced from 55 SaaSHub category listings. It is not a survey and not a random sample of all B2B software companies. See the methodology page for exactly how the sample was built, what the heuristics can and cannot detect, and the known limitations. This page will not be re-crawled on a fixed schedule; a future re-run will be published and dated as its own version rather than silently overwriting this one. Last updated 2026-08-21.
Further reading
Curious where you land in this dataset? Run your own case study library through the same checks we ran on these 1,000 companies.
Audit your own proof against the index →Don't become one of the 39%.
Your first voice interview is free. No credit card required.