Traffic is falling everywhere. Your leads do not have to.

One comparison in a client’s analytics this year told the story better than any industry report. Their website traffic declined 16% against the prior period. Their lead volume held flat, and measured conversion efficiency rose 17 to 20%.

There are two explanations for a gap like that, and the honest answer is that both are probably at work. The first is composition: when AI summaries absorb clicks, they disproportionately absorb informational, top-of-funnel clicks, so the traffic that disappears is the traffic that was least likely to convert, and conversion rates rise partly on their own. The second is architecture: a platform measured and structured well enough to turn a higher share of its visitors into enquiries. Distinguishing between the two requires measurement most organisations do not have configured, and that, more than either explanation, is the point of this article. When the discovery landscape is being repriced, the organisations that can read their own funnel keep their demand. The ones that cannot are flying blind through the repricing.

We think that comparison is the most useful lens for what marketing leaders are facing right now, so this article works through what we are seeing across the properties we manage and audit, what it suggests, and what to check in your own analytics tonight.

The ground moved under everyone

We covered the discovery shift in detail in our analysis of AI-driven discovery, so the short version will do here. Organic click-through is compressing: Pew Research found users click a traditional result roughly half as often when an AI summary appears, and a randomised field experiment measured a 38% reduction in organic clicks on affected queries. AI summaries are one force among several, alongside algorithm changes and the steady fragmentation of where people spend attention, and we are careful not to attribute any single client’s numbers to any single cause. What we can say is that our own client data sits inside the published picture, with declines in exposed sectors landing within benchmark ranges. None of it is an SEO story; the organisations involved run capable search programmes. The environment repriced itself underneath them.

For this article, that shift is context, not the subject. The commercial question it forces is the subject: when the volume of discovery becomes less predictable, how does an organisation protect its demand? Our answer, from the platforms we manage and audit, is that demand protection lives in two places the discovery landscape cannot touch: measurement and conversion architecture.

The demand leak

That question is where our audits keep finding the real problem, and it predates AI entirely.

Across the platforms we review, the same conditions recur. A flagship category page drawing tens of thousands of views in a single period while converting at roughly a third of the site’s own average: a page that informs but never asks, on the exact traffic the business paid hardest to earn. Demand signals firing, form starts, calls, enquiry pages, with none of them set up as key events, so they appear in no report and can be attributed to no channel. Revenue values sitting at zero in analytics, so every budget conversation happens without a dollar figure. In one portfolio, organic search represented 67 to 85% of sessions with effectively nothing behind it: a single, efficient, fragile channel carrying the whole demand load.

Here is the uncomfortable arithmetic those conditions create. When traffic was abundant, a leaking funnel was an invisible tax; volume covered it. As discovery volume compresses, the same leak becomes the difference between a flat quarter and a bad one. And unmeasured demand carries a double cost: some of it genuinely leaks through broken journeys and missing capture, while the rest converts invisibly, unattributable to any channel or decision. Either way, the organisation cannot manage it, defend it in a budget conversation, or notice when it starts to fall.

There is a second reason the leak costs more than it used to: as our discovery analysis suggested, the visitors still arriving tend to be fewer and further along in their decision, so every point of leakage now loses higher-value demand than it did a year ago.

Four signs to check in your own analytics tonight

None of this requires a consultant to diagnose. Four checks, thirty minutes, your own GA4:

1. Are your real conversion actions configured as key events? Key events are GA4’s term for the actions that matter to the business. Open the list. If form submissions, calls, or enquiry actions are missing, demand is arriving unmeasured.
2. Do your conversions carry values? If revenue or lead values sit at zero, every optimisation and budget debate is happening without a dollar figure attached.
3. Where does your highest-traffic content send people? Find your top five pages by traffic and follow the journey. If the answer is nowhere, that is the leak.
4. What percentage of your sessions come from one channel? If a single channel carries 70% or more of your demand, you are exposed to exactly the shift now underway.

Two or more of these and the honest conclusion is that your platform has a measurement and conversion problem that the traffic environment is now exposing.

Conversion performance is the hedge

The response to a volatile discovery landscape is not, in our experience, to publish harder into a compressing channel. It is to make the platform itself the asset. When we say conversion architecture, we mean something plain: everything that determines whether a visitor becomes an enquiry. Measurement configured so demand is visible. Journeys designed so an informed visitor can act without friction. Pages structured so both people and AI systems can understand what you do. Performance foundations that do not squander the visits you earn. We have written before about platforms built to launch rather than to scale; this is the same discipline viewed from the demand side, and it is the only part of the funnel entirely within your control.

It is worth being clear about the split we typically find. Some of this is configuration: key events, values, and attribution can often be fixed in days, and that work pays back immediately by making everything else visible. Some of it is structural: journeys, page architecture, and performance are platform work, and the right scope depends on what the measurement reveals once it exists. The sequencing matters, and it is always the same order: measure first, then decide. Committing to structural work before the measurement exists is how digital budgets get spent on the wrong problem.

The takeaway

Three things we would put in front of any leadership team right now.

First, separate the traffic conversation from the demand conversation. Traffic is being repriced by forces outside your control; demand capture is not. Report them separately, because a team judged on the first will be punished for the landscape, while a team judged on the second can actually win.

Second, establish the baseline before optimising anything. The four checks above are a baseline exercise, and everything strategic depends on them: you cannot protect demand you cannot see, and you cannot evaluate any response, content, platform, or paid, without knowing what today’s funnel actually converts. This is also a governance point. Boards and executive teams are increasingly being asked to fund responses to a discovery shift they cannot see in their own reporting; measurement maturity is what turns that conversation from speculation into a decision.

Third, treat conversion efficiency as the metric that matters most in a volatile discovery environment, and measure it honestly. The client whose traffic fell 16% while leads held flat benefited from some combination of higher-intent traffic mix and a platform that converted what arrived; the organisations that thrive will be the ones who can tell those two apart in their own data, because only one of them is within your control, and it is the one worth investing in.