Traffic, conversion rate and value: the arithmetic that governs everything
What a corporate website produces is three numbers multiplied together: how many of the right people arrive, what share of them do something you can follow up, and what a followed-up enquiry is worth. Traffic times conversion rate times value equals result. Doubling any one term doubles the outcome, so the question for a fixed budget is not which term matters most but which is cheapest to move.
It is rarely traffic. Traffic is bought in an auction against competitors who want it too, or earned over quarters, and paid for again every month. Conversion rate is changed by rewriting a page, removing four fields from a form, or publishing what the product costs in structural terms — one-off costs that keep paying. Value is the most neglected term: a site producing a hundred enquiries a month of which sales dismisses ninety has a value problem, and more traffic only multiplies the wrong number. As a rough test, if the site converts a very small fraction of one percent of visitors into any enquiry at all, traffic is not the constraint.
Four jobs, and most corporate sites are built for two
A B2B website has to do four things. It has to be found, which is what architecture, on-page work and performance are for. It has to be credible to somebody whose own standing is at risk if they recommend the wrong supplier. It has to be forwardable inside the buyer's organisation, because the person reading it is rarely the person who signs. And it has to be easy to buy from: pricing structure, implementation reality, security and contract documentation, and a way to reach a human.
The third job is the one almost nobody designs for. Gartner's May 2025 research, from a survey of 632 B2B buyers fielded in August and September 2024, describes buying teams of five to sixteen people across as many as four functions, and reported that around 74 percent show what Gartner calls unhealthy conflict. It also found group-level relevance improved consensus by 20 percent while individual-level relevance reduced it by 59 percent. Forwardable means a stable URL, a page that states the thing plainly, a document that does not demand a form the third time it is opened, and a security page an IT reviewer can read unaided.
Most sites do the first two jobs and stop. The design review asks whether the site looks like a serious company and whether it ranks; nobody asks where the finance director's question is answered.
Why the site is structurally the hardest thing a marketing team owns
Every other channel has one owner. The website has four constituencies and no owner: marketing wants the change, IT deploys it on a release cycle set by other priorities, sales judges the result in a meeting nobody from marketing attends, and legal reviews anything touching data collection or a product claim. The predictable results: a copy correction that takes six weeks, a content management system chosen on security grounds by people who will never publish in it, a privacy policy describing less tracking than the site performs, and a tracking plan that is nobody's job.
The fix is procedural. Agree in writing which changes marketing may ship without a ticket — copy, form fields, page order, meta descriptions — and which need review, meaning anything that changes what data is collected, what is claimed, or what is promised. Name an owner and a response time for each. Keep one register of every tag and what it collects: it is what legal needs, what the privacy notice must match, and the artefact almost no corporate site maintains.
The diagnostic order: start by distrusting the measurement
Before concluding anything about a website, establish whether the numbers describing it can be trusted. Duplicate tags, unfiltered internal traffic, form events firing on validation errors, thank-you pages counted twice, broken cross-domain tracking, and bot traffic all produce confident charts that are wrong. Two GA4 behaviours bite low-volume B2B sites in particular: data thresholds are system-defined, cannot be adjusted, and Google does not publish the numeric threshold, so rows disappear from small reports; and behavioural modelling requires minimum daily volumes of both consented and unconsented traffic that most B2B sites never reach.
Then work through the site in order. Is the right traffic arriving, judged by query and source segment rather than in aggregate? Do the pages receiving it offer anything worth converting to? Does the conversion mechanism work when tested on a phone, on a slow connection, with a corporate email address a validation script might reject? Is what happens after a submission fast and human? Is performance a real obstacle, measured in the field rather than the lab? The Core Web Vitals are LCP within 2.5 seconds, INP at 200 milliseconds or less and CLS at 0.1 or less, at the 75th percentile; Lighthouse cannot measure INP at all, so a perfect lab score does not mean a site passes. Design comes last.
Redesign is the answer far less often than it is the proposal
Some problems do require rebuilding: an information architecture that cannot express the product line, a template system that cannot produce the page types the buying process needs, a platform that cannot be made fast or accessible, a rebrand or domain change already committed to.
Most problems are not structural. Copy that describes the company instead of the buyer's problem. A form asking for company size, budget and timeline before the visitor has read anything. No pricing information of any kind. Case studies with adjectives and no numbers. A contact page with a form and no telephone number. Each can be fixed this quarter, and each moves the middle term of the arithmetic.
Redesigns are attractive because they are fundable, visible and easy to explain to a board. They also reset whatever was quietly working, put search performance at risk during migration, and absorb a year. Write the fixable list first and work through it: you will understand what a new site needs far better afterwards, and often find you no longer need one.
The profiles here divide three ways: changes available now (conversion rate optimisation, landing pages, lead capture forms, on-page optimisation), structural work (design, architecture, content strategy, migration), and whether you can see any of it (analytics, visitor tracking, reporting, performance, privacy policies). Each says where its discipline is the wrong investment — formal conversion testing on a site without the traffic to reach significance being the commonest.