Your sessions chart looks much the same as it did six months ago. Your rankings have held. The enquiry inbox has not.
Website traffic but no enquiries is the most common symptom owners bring to us, and it is rarely one cause. It is usually one of six, and five of them have nothing to do with AI.
Below is the order we check them in, what each one looks like in the data you already have, and the one cause that leaves no trace in Analytics at all.
Key takeaways
- Traffic and enquiries are separate metrics and they move independently, so a flat sessions line tells you almost nothing on its own.
- Check the boring causes first. Broken forms and broken tracking account for more of these cases than anything else, and both are free to rule out.
- Total sessions can hold while the mix underneath them changes from commercial to informational. The top-line number hides it.
- SparkToro, using Similarweb clickstream data published in June 2026, put the UK zero-click rate at 69.5%, the highest of the six countries measured.
- One cause produces no session, no impression and no referrer. The only way to see it is to ask the question your buyer asks and read the answer.
What does steady traffic with falling enquiries actually mean?
It means the volume of visits held while the quality, the intent or the measurement of those visits changed. Traffic and enquiries are different metrics and they move independently. Six causes account for almost all of it, and they are worth checking in order of likelihood rather than in order of how interesting they are.
That order matters commercially. The first three cost nothing to investigate and can often be fixed the same week. The last three take longer and cost more, so it is expensive to start there and find out later that a form had been silently failing since a plugin update.
Cause one: is your form or your tracking broken?
Start here. It is both the most common cause and the cheapest to eliminate. A plugin update, an expired mail connection, a consent banner blocking the tag, a rebuilt page that lost its tracking code. Enquiries either stopped arriving or stopped being counted, and in a report those two look identical.
Test it properly rather than from your own desk. Submit the form from a phone on mobile data, in a private window, with a real email address you can check. Then confirm the message reached the inbox, the spam folder and the CRM, and that the thank-you page still loads.
Then compare counts. Take the number of key events GA4 recorded last month and put it next to the number of enquiries actually sitting in your inbox for the same period. If those two numbers disagree, you have a measurement problem rather than a marketing one, and the fix is an afternoon.
Cause two: is this seasonal?
Compare the same month last year, not the last three months. B2B enquiry volume reliably thins around August, the second half of December and the fortnight either side of Easter. In tender-led sectors it moves with procurement cycles rather than the calendar, which can put the quiet period almost anywhere.
One caution. A year-on-year comparison assumes last year was normal. If last year contained a site migration, a rebrand or a paid campaign you have since switched off, the comparison is not clean and you need to go back two years.
Cause three: is a competitor bidding on your brand name?
Someone paying for your company name sits above your own result and takes the click before it reaches you. Your organic sessions barely move, because most people still scroll to you or come back later. Your enquiries move, because some of them do not.
Search your own business name in a private window, on a phone, and look at what sits above your listing. It takes thirty seconds and people rarely do it.
Then check branded terms in Search Console. If impressions on your own name are steady but clicks have slipped, something is intercepting them. A defensive bid on your own brand name is usually cheap, because nobody else has a quality score on it.
Cause four: has a ranking or content change moved your traffic mix?
Total sessions can hold while the mix underneath changes completely. You lose a service page and gain a blog post. The count stays flat, the intent collapses. This is the most commonly missed cause, because the top-line number is reassuring and the damage sits one level down.
Segment the landing page report in GA4 and split your commercial pages from your informational ones. Then compare clicks per page, year on year, in Search Console. A service page down 40% inside a flat total is a serious problem wearing a calm number.
Clicks are also being squeezed on queries where an AI summary appears above the results. Pew Research Center, in July 2025, found US users clicked a link on 8% of searches where an AI summary was present, against 15% where it was not. Google publicly disputed the methodology, so treat the size of that gap as contested and the direction as not.
Cause five: are your buyers researching somewhere you cannot see?
Buyers increasingly build a shortlist before they visit anyone. They ask an assistant, read a comparison, check reviews, then arrive at your site already decided, or never arrive at all. Your traffic looks stable because the people who were always going to find you still do. The people forming the shortlist are doing it elsewhere.
The scale of this is measurable even if your share of it is not. SparkToro, working with Similarweb clickstream data published in June 2026, found that 69.5% of UK Google searches end without a click, the highest rate of the six countries in that study. Around 232 clicks reach the open web per 1,000 UK searches.
MapQuest is the useful comparison here, and not for the reason people usually give. The product did not get worse. The map simply arrived somewhere else, inside the phone, and the reason to visit went with it. Nothing about the service changed. The route to finding it did.
Cause six: are you being read but never named?
This is the one that leaves no trace. An assistant reads your website, answers your buyer’s question, names three businesses, and yours is not among them. No session, no impression, no referrer, no line in any report you own. You are ranked but not recommended, and nothing in your Analytics will tell you.
You can check it manually and you should. Write down the ten questions a buyer actually asks before they enquire. Not your keywords, their questions. Then run each one in ChatGPT, Perplexity and Google’s AI mode, and write down which businesses get named.
Be honest about what that tells you. No platform publishes how it selects the businesses it names, so this is observation rather than confirmation, and results vary by wording, location and day. Nobody can promise you a citation, and anyone who does is selling you the same thing the last agency did.
What it does give you is a baseline. Run the free AI visibility check and it does the same work across the major assistants in a few minutes, and tells you who is being named in your place.
What can and cannot be seen in GA4?
GA4 shows you sessions and events. It does not show you interception, intent shift or absence. Three of the six causes above are invisible in it, which is why a flat Analytics dashboard is often taken as proof that nothing is wrong.
| Signal | Visible in GA4 | Where to look instead |
|---|---|---|
| Form submissions | Yes, if the event fires | Your inbox and CRM, counted against GA4 |
| A competitor bidding on your brand | No | Search Console branded queries, Google Ads auction insights |
| Traffic mix shifting to informational | Yes, with segmentation | Landing page report, split commercial from blog |
| AI referrals carrying a referrer | Partly | The AI assistants channel added during 2026. Treat it as a floor |
| AI referrals carrying no referrer | No | Rising Direct alongside falling Organic |
| Being read and never named | No | Manual prompt testing, or a visibility check |
That last row is the important one. Many assistant referrals arrive with no referrer at all and land in Direct, so Organic falling while Direct climbs is worth investigating rather than celebrating.
How do I check whether my forms are broken?
Submit your own form from a phone on mobile data, in a private window, using an address you can check. Confirm it reached the inbox, the spam folder and the CRM. Then compare the number of key events GA4 recorded last month against the number of real enquiries you received. If the two disagree, fix the tracking before you change anything else.
Can I see AI referrals in GA4?
Partly. GA4 added an AI assistants channel during 2026, but it only catches sessions that arrive with a referrer it recognises. A large share arrive with none and are recorded as Direct. Read the channel as a minimum rather than a total, and watch for Direct rising while Organic falls, which is the same traffic wearing a different label.
How long should I wait before acting?
Rule out the free causes immediately, because a broken form costs you enquiries every day it stays broken. For anything involving rankings or content, use a rolling 28-day comparison against the same period last year and give it at least two of those cycles before drawing conclusions. Reacting to three weeks of data usually means reacting to noise.
Is this seasonal in B2B?
Often, yes. August, late December and the fortnight around Easter reliably suppress B2B enquiry volume, and tender-led sectors follow procurement cycles rather than the calendar. Compare the same month last year rather than last quarter. If the shape matches and nothing else changed, seasonality is the likely answer and no intervention is needed.
Where to start
Work through the six in order. Most businesses find their answer in the first three, and it costs nothing to look.
If you get to the end and everything checks out, the cause is the one that does not appear in your reports, and that is a measurable thing rather than a theory. AI search optimisation exists to change what the assistants say about you, and it begins with knowing what they currently say.
We have run this inside facilities management and property rather than read about it. One organic channel went from 40 to over 1,000 monthly visits on commercial terms inside a month, and from no organic bookings to around 10 booked appointments a week. These results were achieved by Marcus Lobow, some in-house as a senior marketer and some through Lobo Digital.
Run the free AI visibility check at lobodigital.co.uk/tools/ai-visibility-checker. You keep the findings either way. If you would rather talk it through first, email hello@lobodigital.co.uk.
