Ask most B2B companies how their website is performing and you’ll get a traffic number. Visits are up, or visits are down, and the conversation ends there.
Traffic is the easiest number to find and one of the least useful on its own. A site can double its visitors and produce fewer enquiries. It can lose a third of its traffic and generate more revenue, if the third it lost was never going to buy anything.
The useful question is narrower. Is the site producing the business outcomes you built it for, and if not, where is it losing people?
Three layers, in order
Website measurement makes more sense when you separate it into layers and read them in sequence.
- Visibility. Are the right people finding you? Impressions, average position, and which queries you appear for.
- Engagement. Do they stay and read? Engagement rate, time on the pages that matter, and which pages they visit next.
- Outcomes. Do they do the thing? Enquiries, quote requests, calls booked, and eventually revenue.
The order matters when something goes wrong. Falling enquiries with steady visibility and engagement points at your conversion paths. Falling enquiries with collapsing visibility is a search problem, and no amount of button-colour testing will fix it.
Visibility
- Are the right people finding you?
- Impressions and average position
- Which queries you appear for
Engagement
- Do they stay and read?
- Engagement rate
- Where they go next
Outcomes
- Do they do the thing?
- Enquiries and their quality
- Revenue, eventually
What to actually track
For a typical B2B marketing site, this list covers what you need. Everything else is optional detail.
- Enquiries, by source. The single most important number, split by how people arrived. Organic search, direct, referral, and paid behave very differently.
- Lead quality. Not every enquiry is equal. Track how many became real conversations, because a page producing twenty tyre-kickers is worse than one producing three genuine prospects.
- Conversion rate on key pages. What proportion of visitors to your service pages take the next step. This tells you which pages are working harder than others.
- Organic impressions and average position, from Search Console. Leading indicators, because they move before traffic does.
- Engagement rate on your important pages. In Google Analytics 4 this replaced bounce rate, and it counts a session as engaged if it lasts over ten seconds, includes two or more page views, or triggers a key event.
- Top landing pages. Where people actually enter the site, which is often not the homepage and frequently surprises people.
- Mobile share and mobile conversion rate, tracked separately. A site that converts well on desktop and badly on phones is a common and expensive pattern.
- Core Web Vitals, from Search Console, as a health check rather than a weekly obsession.
Notice that the first two are business numbers rather than website numbers. That’s deliberate. The website’s job is generating leads worth having, and measurement should be anchored there.
The Eight to Track
Set up the tracking properly, once
Most measurement problems are collection problems rather than analysis problems. Three things are worth doing carefully at the start, because everything downstream depends on them.
Define your key events. In GA4 these are the actions that matter, form submissions, quote requests, calls booked, and file downloads. If they aren’t configured, you have no outcome layer at all and no amount of reporting will produce one.
Separate the enquiry types. A newsletter signup and a quote request should not be the same event. Rolling them together produces a conversion number that goes up while your pipeline stays empty.
Connect the form to something that remembers. A form that emails an inbox is a lead you can’t trace. A form that writes to your CRM is a lead you can follow to revenue, which is the difference between reporting and knowing.
None of this is a big project. It’s an afternoon that determines whether the next twelve months of reporting mean anything.
The metrics that mislead
Some numbers get reported precisely because they’re easy to produce, not because they inform a decision.
- Total pageviews without segmentation. A number that rises when anything happens, including things you don’t want.
- Time on site as a headline. Longer isn’t better if people are lost. A visitor who finds the answer in forty seconds and books a call had a better experience than one who wandered for six minutes.
- Bounce rate compared to old numbers. GA4 defines it as the inverse of engagement rate, which is a different calculation from the old Universal Analytics version. Comparing the two directly produces nonsense.
- Keyword rankings for a single term, checked from your own browser, which is personalized and location-biased.
- Social followers, unless social is genuinely a channel for you.
None of these are useless. They’re just not the ones to lead with, and they have a habit of filling reports that never mention whether the business got anything.
Numbers That Mislead
On benchmarks
You’ll find industry benchmark figures for most of these. Treat them carefully.
For engagement rate in particular, no authoritative industry benchmarks exist for GA4 as things stand, and the numbers circulating vary widely depending on who compiled them and from what sample. B2B sites also differ enormously by sector, deal size, and traffic mix.
Your own baseline is far more reliable. Measure where you are now, then compare against yourself in three months. A site improving against its own history is doing better than one matching somebody else’s average.
A monthly review that takes twenty minutes
Reporting expands to fill the time available, so it helps to fix the shape in advance. Once a month, answer five questions.
- How many enquiries did we get, and from which sources?
- How many were worth having?
- Which pages produced them?
- Is organic visibility trending up, flat, or down?
- What is the one thing we’ll change before next month?
That last question is the one that makes the exercise worth doing. A report nobody acts on is an expensive habit, and most website reporting fails at exactly that point.
Where possible, connect enquiries to what happened next. If your CRM can tell you which web leads became customers, you have the only metric that settles arguments, which is what the website returned against what it cost.
What to do when the numbers are bad
Read the layers in order and the diagnosis usually presents itself.
If visibility is falling, look at search first. Something has changed in rankings, competitors, or the site’s technical health, and our guide to B2B SEO covers where to look.
If visibility holds but engagement is poor, the problem is usually the page itself. Wrong message, wrong audience, or slow loading.
If engagement is fine but enquiries aren’t coming, look at the paths. Unclear next steps, forms asking too much, or proof missing at the moment people decide. That’s usually a buyer journey problem rather than a traffic one.
A worked example
A manufacturer’s monthly report shows traffic down 18% year over year. The reflex is alarm. Read it in layers instead.
Visibility is up. Impressions have grown and average position has improved for the terms tied to their actual products. What fell was traffic to three old blog posts about a topic they no longer sell into, which had been inflating the total for years.
Engagement is up too. Fewer visitors, but the ones arriving now read more pages and spend longer on the service pages.
And outcomes are up 40%. Enquiries have grown, and the sales team says the quality is noticeably better.
The site lost the wrong traffic and gained the right traffic, and the headline number described that as a decline. A report that stopped at traffic would have prompted somebody to fix a site that had just started working.
The inverse happens too. Traffic climbs, everyone is pleased, and enquiries stay flat because the growth came from a piece of content that attracts students and job seekers. Both situations look identical at the top layer and completely different one layer down.
Fewer numbers, better questions
Most B2B companies don’t need more data. They need a shorter list of numbers they actually look at, tied to outcomes they actually care about, reviewed on a schedule they actually keep.
Eight metrics, five questions, once a month. That’s enough to know whether your website is working, and enough to know what to do about it when it isn’t.
If you’d like help working out what your current numbers are telling you, book a discovery call and we’ll look at them together – no pitch, and no pressure.
