61% of B2B buyers would rather buy without a sales rep at all, and 69% say the information on a vendor's website does not match what its sellers tell them. Both figures come from a Gartner survey of 632 B2B buyers, published in June 2025. So a B2B SaaS website has to carry most of the sale alone, and say what the sales team says.
This guide covers which pages the site needs, the order in which every page makes its argument, and how to get to launch in four weeks. Conversion benchmarks are in our B2B SaaS Website Conversion Rate Statistics 2026 post; this one is about structure and process.
What a B2B SaaS website has to do before anyone talks to sales
Buyers arrive late, with a shortlist. 6sense's 2025 B2B Buyer Experience Report, based on nearly 4,000 buyers, puts the point of first contact at 61% of the way through the buying journey. Earlier editions put it at 69%. Buyers fill about four of five shortlist places on day one, and the eventual winner was on that day-one list 95% of the time. 6sense sells intent data, so this is a vendor source, but Gartner and Forrester point the same way.
They prefer to do it alone, up to a point. Gartner found buyers favour self-service for general information but want a seller to judge fit for their company. That is a useful division of labour: the site explains the category, the product and the price logic, and the demo call handles fit.
A committee reads the site, not a person. Forrester's State of Business Buying 2024 puts the average buying decision at 13 people, with most purchases involving two or more departments. Among the self-guided sources buyers used, generative AI ranked first, with internet searches and vendor websites close behind.
NN/g's distinction between users and choosers is the practical way to plan for this. Users care about specifications, daily experience and support; choosers care about cost, reliability, integration effort and return on investment.
The page map: twelve page types and the job of each
A B2B SaaS site needs the right page types more than many pages. This is the map we would start from for a sales-assisted product; a product-led company swaps the demo page for sign-up.
| Page type | Its job | Main reader | At launch? |
|---|---|---|---|
| Homepage | State category and problem, route each reader | Everyone | Yes |
| Product overview | Show how the product works end to end | Users | Yes |
| Capability pages, 2 to 4 | One page per job the product does | Users | Yes |
| Use case or role pages | Restate the problem in one team's words | Users and internal champions | One or two |
| Industry pages | Show fit for a regulated or specialised sector | Choosers | Only if sales sells by sector |
| Integrations | What it connects to, and how | Technical evaluators | Yes |
| Security and trust | Compliance, data handling, uptime, subprocessors | Security and procurement | Yes |
| Pricing | Tiers, price logic or a typical-scenario price | Choosers | Yes |
| Customers | Logos, case studies, named quotes | Choosers | Yes, logos at minimum |
| Comparison or alternatives | Answer "how is this different from X" | Late-stage evaluators | Within the first quarter |
| Demo or contact | Book the call and set expectations for it | Ready buyers | Yes |
| Resources and blog | Search traffic and early research | Early researchers | After launch |
Source: our recommendation, drawing on NN/g's B2B guidance and HockeyStack's 2025 page data
Pricing goes in the main navigation. HockeyStack's December 2025 study of 31 million visitors to 80 B2B SaaS sites found pricing pages take 16.5% of site traffic, demo pages 0.91% and case study pages 0.76%. The report calls the classic homepage, product, case studies, pricing, demo path "broken". The page people go looking for should be one click from everywhere. Google's site-structure guidance, written for shops, notes that the links needed to reach a page, and the links pointing to it, signal how important it is.
Integrations and security get their own pages. NN/g lists integration, compatibility and compliance information as one of five ways B2B sites differ from consumer sites, and calls failing to find it a major pain point for business customers.
Segment navigation must be unambiguous. If the site splits by role, company size or sector, NN/g advises that each segment be clearly defined and mutually exclusive. There is more on how this applies to software companies on our B2B SaaS industry page.
Customer proof does two jobs. On sites without a published price, HockeyStack found visitors appear to use logos and case studies to estimate what the product costs. Visitors who read a case study before submitting a form were 18% more likely to become an opportunity. Launch with a logo row at least, and aim for one case study per main segment.
The message hierarchy: category, problem, outcome, proof
The page map decides where readers go; the message hierarchy decides what they find. We recommend four layers, in a fixed order, on every page that sells.
| Layer | The reader's question | Where it lives | The test |
|---|---|---|---|
| Category | What kind of thing is this? | Homepage headline, title tags, first line of product pages | A stranger can name two competitors |
| Problem | Is this about my situation? | Subhead and first section | A prospect would say it in a meeting, unprompted |
| Outcome | What changes if we buy it? | Product and capability pages, pricing tiers | It describes a result, not a feature |
| Proof | Why should I believe you? | Beside each claim; customers and security pages | A named customer, a standard or a sourced number |
Source: our recommended framework
Category first. April Dunford, the positioning consultant, describes a market category as a way of setting the context for a product. It sets off a set of assumptions in the buyer's mind about what the product does and whom it competes with, and it does not replace the pitch. So it comes first and stays short. "Contract management software for in-house legal teams" lets a reader fill in a dozen facts the page no longer has to state. A coined category forces the page to explain from zero.
The ten-second window. NN/g's write-up of a Microsoft Research study of 205,873 web pages found that users often leave within 10 to 20 seconds, that the first 10 seconds carry the highest probability of leaving, and that the curve flattens at around 30 seconds. NN/g's advice is to communicate the value proposition within 10 seconds. The data dates from 2010, so treat the seconds loosely: category and problem sit above the fold.
Problem in the buyer's words. The problem line works when a prospect could have said it on the discovery call. Call recordings and support tickets are the best source; the investor deck, which describes the market, is the worst.
Outcome before feature. The outcome layer says what is different after the purchase: fewer manual steps, a faster month-end close, an audit that passes. Each capability page should open with its outcome for the choosers, then give the specification for the users.
Proof next to the claim. A testimonial carousel at the foot of the homepage is proof in the wrong place. Put the customer logo beside the claim it supports, the compliance standard beside the data-handling sentence, and the integration list beside "works with your stack".
One document, shared with sales. Gartner's finding that 69% of buyers see inconsistencies between a vendor's website and its sellers is a process failure, not a copy failure. Write the four layers once, in a message document that sales signs off, and draw every page and every deck from it.
Pricing pages when the price depends on the deal
NN/g reported in 2013 that business customers name pricing as the information they most need online, and that test participants went to competitors' sites when a site hid it. When the price depends on seats, volume or configuration, its fix is prices for typical scenarios or a range, not a complex calculator, which its studies found error-prone.
For SaaS that gives three options: tiers with a "from" price and a custom enterprise tier; the price logic (per seat, per unit, platform fee) with one worked example; or, at minimum, what drives the price and what sales needs to quote it. HockeyStack's data shows the trade-off: hidden pricing collects more submissions, published pricing turns more of them into pipeline. Either way, the page exists, sits in the navigation and says something true.
The four-week build, week by week
We build B2B marketing sites on a four-week process that starts with a Strategy Week, in Webflow for the marketing site and in Next.js when the job is an application. That is our bias: four weeks assumes a Webflow build of roughly the page map above, and a site with application features inside it takes longer. We use this sequence on marketing-site projects, including USLege, a legal-tech redesign, and the order matters more than the week count.
| Week | What gets decided or built | Sign-off at the end | What usually slips |
|---|---|---|---|
| 1: Strategy Week | Page map, message document, URL and redirect map, analytics events | Page map and message, signed by marketing and sales | The message, when sales is absent |
| 2: Design | Homepage, one capability page, pricing, one case study; components | Four layouts and the component set | Homepage hero reviews |
| 3: Build | All templates, CMS collections, real content | Staging site with real content | Case studies, security details, integration lists |
| 4: Launch | QA, redirects, tracking, performance, go live | Launch checklist | Legal review of the security page |
Source: Octo Agency four-week marketing-site process; the detail within each week is our recommendation
Week 1, Strategy Week. Nothing is designed until the page map and message document exist. The week ends with two sign-offs: marketing on the page map, sales on the message. The second prevents the inconsistency Gartner measured. If the site replaces an existing one, the URL inventory and redirect map start here, not in launch week.
Week 2, design. Design four representative pages: the homepage, one capability page, pricing and a case study. Between them they contain nearly every component the site needs. The design choices themselves are the subject of our companion post, B2B Website Design in 2026: The 9 Decisions That Decide Whether It Books Demos.
Week 3, build. Templates and CMS collections go up on staging, and real content goes in. This is where four-week projects slip, and the cause is rarely the build: a case study waiting for customer approval, a security page waiting for the compliance lead, an integration list nobody owns. Ask for all three in week one.
Week 4, launch. QA, tracking and redirects. Google's site-move guidance asks for a mapping of every old URL to a new one, server-side permanent redirects (301 or 308) and redirects kept for at least a year. It warns of temporary ranking fluctuation, and that a medium-sized site can take a few weeks or more for new URLs to show. Our Website Redesign Statistics 2026 post has timelines for longer projects.
What this means for a SaaS team planning a new site
Settle the order before the look. Cut the twelve page types to what you can fill with real content in three weeks, write the four message layers in one document that sales signs, put pricing and security in the navigation, and plan for 13 readers rather than one. Start approvals in week one. Then design four pages, build the rest from their components, and keep the redirects for a year.
The result is a site that says what its sellers say, closing the gap 69% of Gartner's buyers reported. Our marketing websites service page explains how we scope a four-week build.
