Of the 185 B2B SaaS customer stories we read in October, 90.3% carry a result number. Only 16.2% carry all four parts a buyer needs to check it: where the customer started, where they ended, how long it took, and what they did to get there.1
Most teams treat the number as the proof, so when sales asks for a stronger case study, the fix is a bigger number. The number was rarely missing. The start and the timeframe were, and those are the two parts a buyer holds their own situation against.
What follows is the published research on how B2B buyers use case studies, read beside our own count of what 185 customer stories from 37 brands carry, niche by niche, and the same test run on our own case studies.
Highlights
Read together, they describe a format buyers lean on at the exact moment they are checking claims, and a published library that gives them claims they can’t check. The quote is nearly always there. The number is nearly always there. The starting point and the date usually aren’t, and that is the part that turns a happy customer into evidence.
Buyers Reach for Case Studies in the Middle of the Deal
Start with when anyone reads one. Demand Gen Report’s 2024 Content Preferences Benchmark Survey asked B2B buyers which content they rely on at each stage of a purchase. In the middle stage:2
The middle stage is where a buyer has stopped browsing and started comparing. Early on they read blog posts and reports to understand the problem. Late, they want a demo of their own data. In between, they want to see someone like them get a result, which is the case study’s one job.
Buyers trust the format, and they notice when it fails them. UserEvidence’s Evidence Gap surveys of B2B software buyers, sellers and marketers:
Marketers already know this, which is why nearly everyone makes them. CMI found 75% of B2B marketers used case studies or customer stories in the past year, and 53% named them among the content that produced their best results, second only to video at 58%.6 So the format isn’t the question. What goes into it is.
Nine in Ten Stories Have a Number, and Fewer Than Half Say Where It Started
To see what goes into one, we read 185 customer stories on 15 October 2026: the first five on the customer-stories index of 37 B2B SaaS brands, four or five per niche across eight niches, each scored for a named customer, a direct quote, a result number, a measured starting value, a timeframe and a mechanism.1 What the stories carry:
The drop happens in one place. Of the 167 stories with a result number, 87 give no start, so the number floats: faster than what, more than when.1 And it isn’t that nobody knew the before state. In 99 of 185 stories it’s described in words, the spreadsheets, the manual triage, the team that couldn’t keep up, with no number attached.1
Buyers notice. In UserEvidence’s first survey, 57% of buyers said their biggest problem with customer evidence was finding credible product and ROI stats, and 50% said finding relevant customer examples.4 A number without a start is the stat that doesn’t read as credible, because the buyer can’t tell whether it describes a team like theirs.
.png)
At Empact Partners, a go-to-market consultancy, customer stories for B2B software companies come out of our Content Marketing workstream, and the interview is built to close this gap. It asks for the measured before state first, then the after, the dates, and the one workflow change that moved it. The quote comes last, because it is easy to get later. The starting number gets harder to find every month after the result lands.
Four Parts Is Where the Pile Thins Out
Put the four parts together and the pile gets small. 16.2% of the stories carry a start, an end, a timeframe and a mechanism somewhere on the page, and that’s generous, because on many of those pages the parts belong to different results.1 Read strictly, asking whether one result is told end to end, 13 of 185 pass.1
What passing looks like is short and plain. Three from the corpus:
None of the three is long. Each gives a buyer a before and an after to set beside their own team.
The share moves a lot by niche. The four-part pass ran from 36% in customer support to 0% in finance and spend, while the share with a result number barely moved, 84% to 96%.1
.png)
That matters more than it looks, because a case study rarely gets read by one person. Gartner’s survey of 632 B2B buyers found buying groups that reach consensus are 2.5 times more likely to report a high-quality deal, and that content relevant to the whole buying group lifted consensus by 20%, while content tailored to one individual cut it by 59%.10
A story with a start and a date is one the champion can forward to finance and the CFO can check. A story with a quote and a percentage only convinces the person who already wanted it.
The Customer’s Voice Is on Every Page, and Buyers Still Check It Elsewhere
The customer’s own words are the part every brand gets right. 93.5% of the stories quote someone at the customer.1 Buyers value the voice, though they weigh it differently from how most teams assume. UserEvidence asked them:
What earns trust is the verification, not the name or the logo. Buyers also go looking for that voice off your site: TrustRadius found 53% of technology buyers spoke to a peer during their buying process, and 74% used reviews to inform their purchase decisions.11
In Demand Gen Report’s survey, 54% of buyers listed content that reads as a sales pitch among their complaints.2 A quote on a vendor’s own page reads, fairly, as the vendor choosing what to show.
The AI engines see it the same way. Qvery, our sister company, measures what ChatGPT and Google AI Mode answer to buyer questions, so weigh what follows accordingly and read the question sets in the articles themselves. Two of its SaaS reads:
Access isn’t the problem. Only 1 of 185 stories sat behind a form.1 The engines can read them. There just isn’t much in most of them that an answer could lift. A dated before-and-after is a fact. A paragraph about a great partnership is not.
Each Niche Proves Itself With a Different Number
The special skill in a case study is picking the number. Across the corpus, time saved led more headlines than anything else, 53 of 185, or 28.6%.1 Money led far fewer: 27 of 185 stories put revenue, pipeline or cost saved in the headline.1
Buyers ask for the opposite. In UserEvidence’s first survey, 67% of buyers said a credible business case around ROI was the most important factor in evaluating software, and 61% said proof of success with customers like them.4
Each niche leads with a different kind of proof:1
| Niche | Leads most often with | Stories | All four parts |
|---|---|---|---|
| Sales tech | Revenue or pipeline | 8 of 20 | 20% |
| Marketing tech | Growth or conversion (revenue or pipeline: 5) | 6 of 20 | 20% |
| Customer support and CX | Productivity per agent | 9 of 25 | 36% |
| Security and identity | Scale (risk or quality: 6) | 7 of 25 | 8% |
| Data and analytics | Time saved | 12 of 25 | 16% |
| Developer tools | Time saved (risk or quality: 5) | 8 of 25 | 12% |
| Finance and spend | Time saved | 9 of 25 | 0% |
| HR and people | Time saved | 6 of 20 | 20% |
The pattern follows who signs. Sales tools lead with pipeline because a sales leader’s number is pipeline. Security vendors lead with scale and risk because a CISO is measured on what didn’t happen. Support leads with output per agent.
Data, developer, finance and HR tools lead with hours, which is honest and the weakest of the lot. Hours turn into money only when the story says how many people saved them, and how often.
G2 found three in four buyers who have lived through a late-stage veto now expect positive ROI within six months of signing.14 Demand Gen Report’s buyers said they want content organized by industry or vertical, 58% of them.2 The buyer is asking “how long until this pays back, for a company like mine”. A case study without a date and a niche answers neither half.
We Ran the Same Test on Our Own Case Studies
The four-part standard is ours. Every result Empact Partners publishes is held to it: a starting value, an ending value, the time between them, and one detail of what produced the change. So we scored the 14 case studies on our own partners page by the same rubric, on the same day.5
Our gaps look like everyone else’s. 14 of 14 name the mechanism, 7 of 14 give a measured start, and 9 of 14 quote a named person at the partner.5 We describe the work in detail and lose the before number, which is the exact failure we just counted in 185 other stories. (Writing about it is less fun than counting other people’s.)
The one that passes cleanly is KDAN. Monthly organic traffic went from 30K to 90K+ within two years, and the page says how: helpful answers published from KDAN team members’ own accounts, and the site and blog localized into more languages.15 Start, end, timeframe and mechanism on one result.
The bottleneck usually isn’t the data. In UserEvidence’s 2025 survey, marketers said customers can quantify the impact 65% of the time, and 47% named getting access to customers as their biggest obstacle.3 36% of marketers had produced two or fewer customer stories in six months.3 The numbers exist. Somebody has to get on a call, ask for them in the right order, and get them signed off.
That’s the part we run. For SystemOne, the customer case studies came out of one-on-one interviews with some of its top customers, inside the same engagement that wrote its core product and solutions pages. In practice, the work runs in this order:
- The audit. Every story you already have, scored against the four parts, so you know which are worth rebuilding and which niche has none.
- The interviews. A call with the customer behind each story, run to get the measured start, the end, the dates and the change that moved it.
- The draft. Written in the customer’s words, with the number and its starting point in the first lines a buyer reads.
- The sign-off. Your customer approves the story, numbers included, before it goes live.
The case-study work sits inside a content program rather than running as a one-off. It is slow, careful work, which is also why it is the part most teams never get to.
The number to watch in any library is the share of stories that give a measured starting value. The corpus sits below half, and more stories with the same gap won’t move it. What our evidence can’t settle is whether a four-part story closes more deals than a quote and a percentage. Nobody measures that, us included. The buyer data says what buyers look for, not what they sign.
If your sales team keeps asking for proof your customer stories don’t give them, book a call with me, and let’s see whether we could work together.
Sources
- Our corpus of B2B SaaS customer stories: the first five story pages on the customer-stories index of 37 B2B SaaS brands in eight niches (185 stories; 3 of 40 brands dropped, index unreadable or merged), each scored for a named customer, a direct quote, a result number, a measured starting value, a timeframe and a mechanism, read 15 October 2026; the one-result read and an audit of the scoring on 16 October 2026.
- Demand Gen Report, “Navigating The Attention Economy Via Snack-able & Shareable Content: The 2024 Content Preferences Benchmark Survey”, 2024; sample size not published.
- UserEvidence, “The Evidence Gap 2025”, survey of 811 B2B software buyers, sellers and marketers, 2025.
- UserEvidence, “The Evidence Gap”, first edition, survey of 619 B2B software buyers, sellers and marketers, 2024.
- Our own case studies: the 14 partner case studies linked from empact.partners/partners, each read in full and scored with the same rubric as our corpus (start, end, timeframe, mechanism, a quote from a named person at the partner), read 15 October 2026.
- Content Marketing Institute and MarketingProfs, “B2B Content Marketing Benchmarks, Budgets, and Trends: Outlook for 2025”, 980 B2B marketers, fielded June to August 2024.
- Sentry, “How Supabase cut weekly crashes by 60x”, customer story, read 15 October 2026.
- Sprout Social, Ivanti case study, read 15 October 2026.
- Zendesk, “Docebo enhances B2B CX with Zendesk AI”, customer story, read 15 October 2026.
- Gartner, “Gartner Sales Survey Finds 74% of B2B Buyer Teams Demonstrate ‘Unhealthy Conflict’ During The Decision Process”, press release, survey of 632 B2B buyers, 2025.
- HG Insights, “Trust More, Verify Everything”, on the TrustRadius 2026 B2B Buying Disconnect Report, 1,862 technology buyers, 2026; with the report’s release, TrustRadius, 15 July 2026.
- Qvery, “How Often AI Answers Cite G2 for SaaS: Mostly on Recommendation Questions”, SaaS questions on ChatGPT and Google AI Mode, September 2026.
- Qvery, “The Landing Page Route: How SaaS Product Pages Become AI Engine Citations”, constrained SaaS questions on ChatGPT and Google AI Mode, August 2026.
- G2, “New G2 Research: AI Is Reshaping How B2B Software Deals Get Made”, release of the 2026 Buyer Behavior Report, 1,000+ B2B software buyers, 2026.
- Empact Partners, KDAN case study, read 15 October 2026.
