.svg)
Some buyer questions are easy to settle alone. Does it work with our CRM? Can I try it? B2B websites gave a full answer to those 69% of the time. Other questions come right before someone signs. What does the SLA promise? How long is the contract? What happens to our data if we leave? Websites gave a full answer to those only 50% of the time.
I run marketing at Storylane, and those numbers come from the Storylane Buyability Index. Between September 8 and October 20, 2026, it scanned 529 B2B websites and checked 10,580 buyer questions against them. Each question got a full answer, a partial answer, a "contact us", or nothing.
Two caveats come first. The questions are generated for each company, based on what a buyer of that product would plausibly ask. They are not logged from real buyers. And the 529 sites were chosen by the people who ran scans, so this is not a random sample of the web.
I call the difference the assurance gap. The more a buyer needs to be sure, the less the website says.
What is the assurance gap?
The Storylane Buyability Index tags every question in two ways. First, who would ask it. That could be an end user, a technical evaluator, security and compliance, procurement and finance, or an executive. Second, how sure that person needs to be.
A self-serve question is one a buyer asks to check fit. A high-assurance question is one they ask before putting their name on a decision. Self-serve questions got a full answer 69.0% of the time (3,293 of 4,770). High-assurance questions got a full answer 49.8% of the time (2,779 of 5,576).
The gap shows up for every type of buyer. Compare each row on its own.
| Who is asking | Self-serve questions fully answered | High-assurance questions fully answered |
|---|---|---|
| End user | 72.9% (of 3,345) | 50.6% (of 1,571) |
| Technical evaluator | 75.1% (of 1,617) | 54.6% (of 2,000) |
| Security and compliance | 78.2% (of 119) | 58.9% (of 2,019) |
| Procurement and finance | 56.7% (of 1,370) | 43.0% (of 2,530) |
| Executive | 63.5% (of 1,893) | 44.5% (of 2,611) |
For every buyer type, the high-stakes questions got fewer full answers, by 14 to 22 points. One question can be tagged with more than one role, so the rows overlap. The self-serve security row is also small.
This is not a few weak sites dragging down the average. 406 sites had at least five questions of each kind. 303 of them answered their high-assurance questions less fully than their self-serve ones.
Which buyer questions do B2B websites answer worst?
The index also groups questions by topic, using keywords in the question. The worst topics are almost all about commitment.
| Topic | Questions (sites) | Fully answered | Partly answered | Nothing found |
|---|---|---|---|---|
| Support SLA and uptime | 235 (232) | 25.1% | 65.5% | 6.4% |
| Contract and cancellation terms | 712 (430) | 39.6% | 39.6% | 12.4% |
| Implementation time | 957 (502) | 43.2% | 49.0% | 7.5% |
| Where data is stored | 360 (356) | 44.4% | 51.7% | 1.9% |
| SOC 2 | 333 (330) | 60.1% | 26.7% | 10.8% |
| SSO and SAML | 220 (220) | 68.6% | 20.5% | 10.9% |
Only one in four SLA questions got a full answer. One in eight contract and cancellation questions found nothing at all. These are not rare questions. Finance and legal ask them before every software purchase, and they came up on hundreds of sites.
Look at SOC 2 and SSO. Those are yes-or-no facts with a badge, and sites publish them. The questions that need a full sentence go missing. What is guaranteed? What is excluded? What happens next?
Why do websites say less when the stakes go up?
The easy answer is that companies hide these answers on purpose. Pricing is negotiable and SLAs are custom, so the site says "talk to us".
The data doesn't really support that. The index counts a fair gate (like a custom SLA, where "talk to us" is normal) and a dodge (a "contact us" where a straight answer was expected) separately. Together they cover just 2.8% of high-assurance questions. What grows is the partial answer. It rises from 25.3% of self-serve questions to 39.8% of high-assurance ones. On SLA and uptime, 65.5% of questions got a partial answer. The site said something, but not enough to rely on.
I have a theory, and it is about who writes the website. Marketing writes it for the person who arrives first. That person asks about fit, and marketing is good at fit. The high-stakes answers belong to other teams. Legal owns contract terms. Customer success owns support promises. Security owns trust documents. Finance owns billing. None of them owns a web page.
So the site gets marketing's summary of their work. "Enterprise-grade reliability." "Flexible plans." "99.9% uptime" with nothing on how it is measured or what happens when it is missed. That is what a partial answer looks like.
Where the full answers do exist tells the same story. For high-assurance questions, 18.0% of full answers came from trust, security or legal pages. For self-serve questions, it was 3.5%. The homepage supplied just 1.2% of high-assurance full answers. The answers that hold up are written by the teams who own the promise. They sit on pages marketing rarely links to.
Who reads the answers your website leaves out?
This is why the gap is expensive. The people asking high-stakes questions are mostly not on your sales calls.
Internal Storylane deal data covers 637 B2B deals and 2,021 recorded sales calls. The average buying account had about 23 known contacts in the CRM (median 10). On average, 2.3 people joined a call. Roughly one in five known contacts ever joined one. The rest got forwarded summaries, or ran the security review and budget approval on their own.
Put those two findings together. The champion who comes to your demo asks fit questions, and your website answers those well. The procurement lead, the security reviewer and the executive who never join a call ask high-stakes questions. Your website answers those worst. They meet you through a forwarded link, and the page they land on says "flexible plans".
We heard a version of this at industry events this month. Several people at our booth said they weren't the decision maker and needed something to share inside their company. I think the website is the biggest thing you have that people can forward. For the questions that decide the deal, it is often the weakest.
How buyers arrive makes this worse. When an AI assistant suggests a vendor, it usually suggests several. The buyer lands on your site with other vendors open in other tabs. If a rival's site states its cancellation terms and yours doesn't, you lose ground without knowing you were compared. You won't see it in your analytics. You will see a deal stall in procurement, or never hear about the deal at all.
How do you close the assurance gap?
The fix is mostly writing.
- List the high-stakes questions. Write down the ten things procurement, security, finance and executives ask before signing. Think SLA and what happens if you miss it, contract length, cancellation and what happens to their data, where data is stored, support response times and implementation time. For each, find the page that answers it fully, in words someone could paste into an internal email. If all you have is an adjective, mark it partial.
- Publish terms, not adjectives. "99.9% uptime" is half an answer. A full answer says how uptime is measured, what is excluded, and what the customer gets when you fall short. If the answer depends on the plan, publish the default and say what changes on the next plan.
- Be honest when you gate. Some things really are custom. Say what the standard position is and what is negotiable. "Our standard SLA is X, and enterprise customers can negotiate Y" is an answer. "Contact sales" on its own is not.
- Put the answers where forwarded readers land. Link the trust, legal and support pages from pricing and product pages. Consider one page for the person who wasn't on the call, explaining in plain words what they are committing to.
- Give one person the job. Legal, customer success and security supply the facts. Someone who owns the website has to turn them into pages and keep them current. Without an owner, the gap comes back the next time a contract template changes.
If you are writing site copy for the whole buying group, our notes on writing a B2B website for a buying committee go further. If security reviews are where your deals stall, the security questionnaire playbook for presales covers the sales-call side of the same problem.
Can a website agent help close the assurance gap?
We sell one, RepX, so weigh what follows with that in mind.
Many full answers already exist on your site. They just sit far from where buyers start, in docs, legal pages and help articles. RepX is built for teams in exactly that spot. You train it on your website, docs, files and sales transcripts. It brings the right answer to the buyer at any hour, instead of making them hunt for it, and hands sales-ready buyers to your team.
It works best when your terms are written down. Do the writing in the list above, and the agent can give every procurement lead and security reviewer the same clear answer, even when they never talk to sales. Buyers also ask what an agent is trained on and what it does with their data. We wrote up what B2B buyers ask about AI agents and privacy.
The one thing to take into Monday
Your website is written for the person who shows up. The deal is decided by people who don't. Pick the five high-stakes questions your buyers ask most. Check whether your site answers each one fully. Fix the ones where all you have is an adjective. If you want to see an agent answer from your real terms on your own site, book a walkthrough with our team.