What is the difference between traffic and traffic with intent?

Traffic with intent is the smaller group who want what you sell and then do something about it: fill the form, ring the number, ask for a price. The two numbers can move in opposite directions, and when they do it's the intent number that tells you what's happening to the business.

Are impressions a vanity metric?

Not on their own. Impressions tell you whether you're visible at all, and for what, which is useful. They turn into a vanity metric the moment somebody reports them as a result. An impression means you were in the running. That's all.

How can enquiries rise when rankings fall?

A few ways. The mix of visitors shifts towards a smaller but better qualified group. Or the channel that fell was never sending anyone in the first place. Or broad head terms drag the average position down whilst the specific pages that convert stay exactly where they were. All three are common and none of them show up in a headline traffic number.

What should a search report actually contain?

Enquiries, calls and form submissions first, split by the channel that produced them. Then the pages and queries behind those actions. Rankings and impressions go further down, as diagnostics that explain the numbers above them. A traffic chart on page one is answering a question nobody in the business asked.

Does this mean traffic doesn't matter?

No, and that would be the wrong lesson to take. Volume is a real constraint. A site with almost no relevant visitors can't convert its way out of that, and both sites below would be better businesses with more of the right people arriving. The argument is about which number tells you whether the work is succeeding.

Should every page be judged on whether it generates enquiries?

No. Some pages exist to convert and some exist to be found, quoted and linked to. A reference page that earns citations and never produces an enquiry is doing its job, as long as somebody decided that was its job in advance instead of finding out from a report six months later.

Why would a site get almost no traffic from Google but still generate leads?

Because Google isn't the only search engine, and for some sites it isn't the main one. Bing powers its own results plus DuckDuckGo, Yahoo, Ecosia and Microsoft Copilot. A site can look close to dead in Google Search Console whilst a different engine family sends most of its visitors and nearly all of its enquiries.

How long does it take to tell whether search work is producing enquiries?

Longer than most reporting cycles allow. Small numbers bounce around month to month, so one good month proves very little. Look at three months minimum, and be wary of anyone drawing a firm conclusion from four weeks of data, us included.

Traffic With Intent: Why Falling Visibility Can Mean Rising Enquiries

In June, one of our client sites had 19,144 impressions in Google and produced no enquiries at all. By August it had 731 impressions, a fall of 96%, and it was producing four enquiries a month plus ten people tapping the phone number.

Impressions, average position, clicks: every number a search report leads with got worse. Enquiries went from zero to four a month.

I have this argument with prospects more often than any other, so here are the full figures.

The numbers

Two client sites, both UK directories, both young. We manage the AI search work on both.

A disclosure first, because it changes how you should read this. These are managed accounts we also hold an interest in.

They’re separate businesses with their own operations, and the search work is real client work.

Being close to them is why I can show this much: month by month enquiry data, straight out of Search Console and Analytics, with no approval process to negotiate and nothing sanded off.

HealthWise, a UK care clinics directory:

Month	Google impressions	Average position	Enquiry forms	Phone and contact taps
June 2026	19,144	32.3	0	2
July 2026	3,230	62.5	2	5
August 2026	731	55.6	4	10

Impressions down 96%. Average position 23 places worse. Enquiries up from nothing, contact taps up five-fold.

Smile Today, a UK dental directory:

Month	Google impressions	Average position	Enquiry forms	Phone and contact taps
June 2026	59,866	25.1	9	2
July 2026	44,022	23.5	7	30
August 2026	56,795	33.0	9	92

Here the average position got eight places worse whilst contact taps went from 2 to 92.

Across the most recent thirty days, which straddles the calendar months above, Smile Today took 997 visits and produced 121 commercial actions. Roughly four a day.

HealthWise took 641 visits and produced 41.

What actually happened

Two different causes, and both are fairly ordinary once you look properly.

HealthWise gets its visitors from Bing. Around 142 a month from Bing and the engines running on its index, which is DuckDuckGo, Yahoo, Ecosia and Microsoft Copilot.

From Google it takes almost none.

So the collapse in Search Console was real, and it described a channel that was sending the site nothing.

Search Console, the dashboard everybody opens first, was measuring an engine the site doesn’t use. It would have reported a disaster every month.

Smile Today is the more common pattern. Its impressions grew because it started showing up for broad national terms it has no realistic chance of ranking for.

Those terms dragged the average position down.

Meanwhile the pages that earn enquiries, individual practice pages and treatment-plus-city pages sitting between positions 5 and 12, carried on doing what they’d always done. The site-wide average got worse because the site got broader.

Where the enquiries come from

Across the accounts we manage with a properly instrumented conversion path, over four weeks:

Channel	Share of visits	Share of enquiries
Organic search	22.8%	96.3%
Direct	65.2%	2.7%
Email	0.5%	1.1%
Referral	9.7%	0.0%

Organic was under a quarter of the visits and 96.3% of the enquiries.

I’m a strong believer that a report should open with the number the business is judged on, and for most firms that is not sessions.

The Direct figure is worth understanding on its own. It’s where bot traffic, mis-tagged visits and anything with a stripped referrer piles up.

It works the other way round too

A site can rank on page one, turn up in AI answers, and still generate nothing.

I see this often in prospect calls. It’s my observation, not something I can put anyone else’s numbers behind.

Businesses arrive with genuinely good visibility. Page one positions on terms they care about. Regular appearances in AI answers. And no enquiries they can trace to any of it.

Nobody has put the reporting next to the sales pipeline. When I do it on a call, that’s usually the first time anyone has.

Not every page is supposed to convert

Taken too far, this argument says every page must produce enquiries and anything that doesn’t is waste. That’s the same mistake, just pointing the other way.

Pages get hired for different jobs. A reference page, a statistics page, a genuinely useful explainer: those earn citations in AI answers, attract links, and build the topical credibility that makes your commercial pages rank at all.

Judging them on enquiry volume is as daft as judging a pricing page on how long people spend reading it.

So decide what job a page was hired to do before you judge it, then measure it against that job. Our guide to measuring AI search visibility sets out the full reporting order.

Traffic still matters

Volume is a genuine constraint. A site with almost no relevant visitors can’t convert its way out of that.

Both sites above would be better businesses with more of the right people turning up.

My point is narrower. A report that leads with visitor counts will eventually tell you something untrue about the business.

How to set the tracking up

Most sites can’t produce the report I’ve described, because the events that represent intent aren’t being recorded anywhere.

That’s more common than it sounds and it’s usually nobody’s fault. The tracking got set up once, at launch, by whoever built the site, and nobody has been back to it since the business changed.

Three things worth checking this week.

Is every route to contacting you firing an event? Not just the main form. Phone taps, email links, WhatsApp buttons, booking calendars.

Booking calendars are the one missed most often. They’re usually embedded from a third party, and the booking completes inside an iframe where your analytics can’t see it.

If your main call to action is “book a call” and your booking tool is embedded, there’s a good chance none of those bookings reach your reporting at all.

Are those events marked as conversions? In GA4 an event does nothing useful until it’s marked as a key event. Plenty of sites are firing perfectly good events that never reach a conversion report because that box was never ticked.

Can you tell which channel produced each one? This is the part that matters most and goes missing most. An enquiry count with no channel attached gives you a total and nothing else.

Everything above depends on being able to say organic search produced 96.3% of the enquiries from 22.8% of the visits. You can’t say that without channel-level attribution.

Channel attribution is rarely anybody’s job, so it doesn’t get done, and the report carries on leading with traffic because traffic is the number that’s already there.

What these numbers don’t prove

I’d rather state these than have somebody find them.

Treat these as one worked example, not a typical outcome. The same rule applies to anybody selling you results: we wrote about how to tell earned authority from manufactured signals.

The windows are short and the numbers are small. Three months of data on HealthWise, five on Smile Today. Four enquiries a month is four enquiries a month.

These figures are big enough to show that two metrics can move in opposite directions. They won’t tell you anything about typical results, and quoting them that way would be the exact thing I’m complaining about here.

Small numbers bounce around, so look at three months minimum before you conclude anything, including from me.

Do this with your own numbers

Open your search report and your enquiry log side by side, same period. For every enquiry, find the page and the query that produced it.

Then look at what your report leads with, and ask whether the two are related.

If you would rather we did it with you, that is what the free AI visibility audit covers. You can also run the AI Visibility Score yourself in about a minute, no email needed.

What did you find?