Pricing ranked first of 28 types of information that business buyers looked for on vendor websites, 29% ahead of the next item, when Nielsen Norman Group tested 179 B2B sites in 2006. Twenty years later the answer has not moved: TrustRadius's 2026 survey of 1,862 technology buyers found transparent pricing at the top of buyers' wish list for the fourth year running.
This guide sets out nine B2B website design decisions that decide whether a site books demos, the evidence behind each, and the test that proves it. The page map and message hierarchy are covered separately in The B2B SaaS Website Playbook: Page Map, Message Hierarchy and the 4-Week Build.
The nine decisions on one page
Unlike most lists of B2B website design best practices, this one keeps only decisions with published evidence and a test you can run.
| # | Decision | Default to start from | Test that proves it |
|---|---|---|---|
| 1 | Pricing visibility | A price, a range or priced scenarios | Opportunities per 1,000 pricing-page visitors |
| 2 | Main call to action | Self-serve action first, demo in the header | Opportunities per 1,000 visitors, one sales cycle |
| 3 | Navigation | Visible on desktop, labelled by topic | Tree test: success and directness |
| 4 | Segmentation | Segments in the buyer's own terms | First-click test per segment |
| 5 | Proof placement | A named result beside every ask | Opportunity rate, proof seen or not |
| 6 | Gating | Gate research and tools only | Gated pipeline against lost organic entrances |
| 7 | Forms | Labelled fields, inline validation | Field-level drop-off and errors |
| 8 | Page speed | Core Web Vitals pass on mobile | Field data, 75th percentile |
| 9 | Accessibility | WCAG 2.2 AA on every template | Scan plus keyboard-only demo form |
Source: Octo Agency, drawing on the research cited in each section
Decision 1: show a price, or at least its shape
In NN/g's 2006 study, users succeeded at 58% of tasks on B2B sites against 66% on consumer sites, and missing prices were the most user-hostile omission. Its 2013 follow-up reported participants leaving for competitors when a site showed no price.
Our B2B SaaS Website Conversion Rate Statistics 2026 post covers HockeyStack's 2025 pricing data; for design, the useful view is per 1,000 visitors. Multiply the form rate by the submit-to-pipeline rate and a transparent pricing page yields about 4.9 opportunities per 1,000 visitors, a hidden one about 4.7. The pipeline is the same. The hidden page simply makes sales process 46 forms to get there, against 28.
| Option | What the buyer sees | Fits when |
|---|---|---|
| Public price list | Exact tiers and seat prices | Standard tiers; 62.5% of HockeyStack's sample |
| Starting price or range | A "from" figure or band | Price varies by volume, floor is stable |
| Priced scenarios | Two or three typical customers and their cost | Services and custom builds |
| "Contact sales" only | No number | Rarely; sales absorbs the screening |
Source: HockeyStack, 2025; Nielsen Norman Group, 2006 and 2013
NN/g also warns against the obvious workaround: in its studies, price configurators were error-prone and only highly engaged customers used them.
The test. Do not judge a pricing change by form submissions; they will fall. Track opportunities per 1,000 pricing-page visitors for a full sales cycle before and after.
Decision 2: lead with what the buyer can do alone
Gartner's survey of 646 B2B buyers in 2025, reported by Digital Commerce 360, found 67% prefer a rep-free buying experience. TrustRadius found 83% of technology buyers shortlisted three or fewer products, with demos, free trials and user reviews among the most influential resources.
In HockeyStack's data the demo page takes 0.91% of site traffic, bounces 70% of it and converts 5.5%. It converts well because almost nobody reaches it, so the real decision is what the rest of the site asks for: a next step without a calendar, such as pricing, an interactive tour, a trial or a specification sheet. How often each appears on SaaS homepages is counted in B2B SaaS Demo Request and Lead Form Statistics 2026.
Our default is a self-serve action as the hero button, with "Book a demo" fixed in the header on every page. For a service business with nothing to try, the self-serve action is the priced-scenarios table from Decision 1.
The test. Count opportunities per 1,000 visitors, not clicks, over at least one sales cycle. Most B2B sites lack the traffic for a split test here, so a dated before-and-after with the same channel mix is the honest version.
Decision 3: navigation that is visible and labelled by topic
NN/g's 2016 study of hidden navigation tested 179 participants on six live sites. Hiding the menu cut content discoverability by more than 20% and made desktop tasks at least 39% slower.
| Measure | Hidden | Visible | Combination |
|---|---|---|---|
| Navigation used, desktop | 27% | 48% | 50% |
| Navigation used, mobile | 57% | Not reported | 86% |
| Seconds to first use, desktop | 31 | 26 | 23 |
| Difficulty, 1 to 7, lower is easier | 2.6 | 2.1 | 2.3 |
Source: Nielsen Norman Group, Hamburger Menus and Hidden Navigation Hurt UX Metrics, 2016
Labels matter too. NN/g's 2014 guidance against format-based navigation applies to the B2B "Resources" menu split into Videos, Webinars, eBooks and Blog: the buyer has to guess which format holds the answer. Organise the top level by topic or problem.
The test. Tree-test the proposed menu before design, with tasks written as buyer questions. NN/g's 2023 guide measures success rate and directness; low directness points at an ambiguous label.
Decision 4: segments buyers recognise in themselves
NN/g's 2006 study listed confusing segmentation among the main B2B failures: divisions that do not match how customers see themselves, and "small business" or "enterprise" sections with no definition. A 400-person company cannot tell whether it is "mid-market". Define each segment in the visitor's terms, such as a headcount, a job title or a problem in their words, and where segments overlap, offer one path per need.
The test. A first-click test with three or four people from each segment: show the homepage, describe their situation in one sentence, record the first click. Nielsen's 2000 analysis found five users uncover about 85% of usability problems.
Decisions 5 and 6: put proof next to the ask, and gate almost nothing
HockeyStack found case study pages draw only 0.76% of site traffic, yet people who read one before submitting were 18% more likely to become an opportunity. TrustRadius's 2026 report found 74% of buyers use reviews to inform a purchase, against 13% who use analyst reports.
So placement decides. Proof on a separate "Customers" page reaches 0.76% of visitors; proof on the pricing page and beside the demo form reaches everyone about to decide. The useful unit is a named customer with one measured result, not a row of logos.
Gating is the other side of the same decision. NN/g's 2015 article says case studies, FAQs and specifications usually do not belong behind a gate, and that search engines usually cannot see gated content.
| Content | Gate it |
|---|---|
| Case studies, specifications, security documents | No: buyers need them to decide |
| Blog posts and guides | No: search cannot index gated pages |
| Original research | Optional, partway through the piece |
| Webinars, calculators, templates | Yes, with one field |
| Trial, quote, consultation | Yes |
Source: Nielsen Norman Group, When to Hide Content Behind Forms and When to Give Content Away, 2015
The test. Record "viewed proof before submitting" in the CRM and compare opportunity rates for the two groups each quarter.
Decision 7: forms that fail gracefully
Field counts are covered in Demo Request Conversion Rates 2026, so this decision is about the failure most audits miss. Baymard Institute's January 2024 research found 31% of sites in its benchmark have no inline validation. The benchmark covers e-commerce checkouts, so treat it as a proxy. Baymard's guidance: do not validate a field before the user has finished it, clear the error once it is fixed, and still show errors on submit.
The second failure is invisible to sighted testers: WebAIM's 2026 scan of the top million home pages found missing form input labels on 51% of them. A placeholder that vanishes on focus is not a label.
The test. Field-level analytics on the demo form: the share of starts reaching each field, and the error rate per field. A high error rate is a design problem; a high drop-off is a question the buyer does not want to answer yet.
Decision 8: a speed budget measured in the field
Google's Core Web Vitals thresholds, last updated on web.dev in May 2025, are assessed at the 75th percentile of real page views, separately for mobile and desktop. The HTTP Archive's 2025 Web Almanac shows how many sites meet them.
| Metric | Good | Poor | Good, mobile | Good, desktop |
|---|---|---|---|---|
| Largest Contentful Paint | 2.5 s or less | Over 4 s | 62% | 74% |
| Interaction to Next Paint | 200 ms or less | Over 500 ms | 77% | 97% |
| Cumulative Layout Shift | 0.1 or less | Over 0.25 | 81% | 72% |
| All three, home pages | n/a | n/a | 45% | 47% |
| All three, inner pages | n/a | n/a | 56% | 61% |
Source: web.dev, 2025; HTTP Archive Web Almanac 2025, July 2025 CrUX data
The home page fails most, and on many B2B sites it carries the hero video, the chat widget and most third-party scripts. Set a budget per template before the hero is designed.
The test. Read field data at the 75th percentile, through Search Console's Core Web Vitals report or the Chrome UX Report, for the home, pricing and demo pages on mobile. A lab score only helps find the cause.
Decision 9: accessibility as a pass or fail gate
WebAIM's February 2026 scan found detectable WCAG 2 failures on 95.9% of the top million home pages, an average of 56.1 errors per page, counting automated checks only. Six error types make up 96% of them, and each can be checked at the design-system stage, so the baseline is cheap to reach.
| Failure | Share of home pages |
|---|---|
| Low contrast text | 83.9% |
| Missing image alt text | 53.1% |
| Missing form input labels | 51% |
| Empty links | 46.3% |
| Empty buttons | 30.6% |
| Missing document language | 13.5% |
Source: WebAIM, The WebAIM Million, 2026
WCAG 2.2, in the W3C Recommendation dated 12 December 2024, adds criteria that bear directly on B2B sites: 2.5.8 Target Size (Minimum) for small header links, 2.4.11 Focus Not Obscured for sticky headers and cookie banners covering the focused field, and 3.3.7 Redundant Entry for multi-step demo forms that ask for the same details twice.
The test. An automated scan of every template, one person completing the demo form with a keyboard only, and a screen reader pass over the pricing table.
How to use this
Write the nine decisions down before anyone opens a design tool, each with an owner and the metric from the first table. They are cheap to change in a template and expensive after launch.
Then test in order: pricing and the main call to action first, because they change pipeline; navigation and segmentation before design starts; proof, gating and forms once there is traffic. Speed and accessibility are gates rather than tests: a template that fails either does not ship.
When you choose a B2B website design agency, ask how it would settle each of the nine and what it would measure. In our four-week marketing-site process the first week is a Strategy Week, which is where we would expect these decisions to be written down before the Webflow build starts. Our marketing website service explains how we scope it.
