On a client call this week, we walked through the kind of number that should worry anyone still treating Google Search as a set-and-forget channel. Cost per sales qualified lead had climbed to nearly six times what it was three months ago. Same product, same landing pages, same campaigns that had run profitably for years.
The account was not the problem. Google was. And the fix we are building with them, micro conversions in Google Ads feeding a Demand Gen strategy, is the same playbook I would hand anyone watching their search volume evaporate right now.
Why your Google search ads are losing volume: AI Mode
Try running a few searches on your phone. Somewhere around the second or third query, Google slides you into its AI Mode experience. The button to get back to normal results is small, grayed out, and tucked up near the logo. Most people never find it. They either take the one or two links the AI serves them, or they bounce.
That is the mechanism. Users who used to scan a results page full of ads and organic listings now get an AI answer with a couple of citations. Fewer real results pages served means fewer ad impressions, and the impressions that survive get more expensive.
The client I mentioned sells a high-ticket consumer product with a long consideration cycle. Their search traffic is down to roughly a seventh of what it was two years ago. It did not fade gradually. It fell off a cliff over the last few months as Google pushed the AI experience deeper into mobile search.
This is not one account's bad luck. Industry analysis of AI search puts 58.5% of US searches ending without a click to any website, and AI Mode accelerates it. The long-tail, high-intent queries that lead gen campaigns are built on are exactly the ones AI answers absorb first.
You cannot fix this with better ad copy. The auction itself is shrinking behind an AI layer. New AI ad surfaces are emerging, and I wrote an honest review of ChatGPT ads for B2B after testing the beta, but none of them are mature enough yet to replace the volume search used to hand you. So the practical question becomes: where does the budget go, and how do you keep the machine learning while it moves?
What micro conversions in Google Ads actually are
A micro conversion is an upstream signal that sits between a click and a lead. Someone views five or more pages in a session. Someone spends more than a minute actively on the site. Someone hits the pricing page or starts a quote. Deeper than traffic, earlier than a form fill.
Why they matter comes down to how bidding works. Smart Bidding learns on conversion volume. A lead gen account producing a handful of qualified leads a month is starving the algorithm, and when search volume drops, it starves faster. Campaigns stall, or worse, they chase whatever cheap clicks are left.
Micro conversions give the algorithm enough signal volume to learn from, if you pick the right ones. And that is a real if.
This is where the Reverse Optimization Trap applies. Feed Google a signal that does not correlate with revenue and it will happily optimize against your business. Scroll depth is the classic offender: optimize for it and Google will find you the world's most committed scrollers. The signals worth using are the ones you have checked against actual pipeline first.
The playbook: shift budget to Demand Gen without going blind
Here is what we are building with this client, step by step.
Step 1: Build engagement conversions in Tag Manager and GA4
Define the events that mean someone is genuinely engaging, not just landing and leaving. The ones we use most:
- Five or more page views in a single session
- Sixty seconds or more of active time on site
- A pricing, configurator, or booking page view
- A meaningful start: first step of a form, quote, or checkout
Build them in Google Tag Manager, send them to GA4, and import them into Google Ads as secondary conversions while you validate that they actually track with quality.
Step 2: Point Demand Gen at the micro conversion
Demand Gen increasingly covers what YouTube, Discovery, and Display used to do separately, and it performs when it optimizes toward a conversion it can actually achieve at volume. Asking it for MQLs it might see twice a week teaches it nothing. Optimizing toward engaged visits tells it something useful: we will pay more for people who behave like buyers, not for the cheapest click in the auction.
Audience construction matters as much as the conversion goal. This client is layering in audiences built from adjacent-intent sites their buyers already use, and feeding those colder audiences the deeper educational content that used to be reserved for retargeting. Meet the intent where it actually lives now, since it is no longer living on a results page.
Step 3: Close the quality loop with identification and enrichment
Micro conversions get you volume. Quality comes from verifying who those engaged visitors are. In B2B, visitor identification tools can match site traffic against IP and firmographic data and tell you whether a visitor fits your ideal client profile. In high-ticket consumer, lead enrichment does the same job: one wealth management client of ours runs every new lead through an enrichment step that scores fit within minutes instead of weeks.
Either way, the move is the same. When a visitor or lead checks out as qualified, send that click ID back to Google as a conversion. Now the algorithm is learning from qualification, not just activity.
Step 4: Judge the whole thing on business outcomes
Demand Gen will look worse than search on a platform dashboard. Click-through rates and cost per click read differently in a feed than in a keyword auction, and comparing them is a waste of a meeting. The questions that matter: what is cost per SQL doing, and what is pipeline doing? That is the scoreboard, and it is the only one.
The tradeoffs nobody puts in the pitch deck
Demand Gen is demand creation, not demand capture. Search caught people who were already looking. Demand Gen has to earn attention first, which means feedback loops run longer and creative matters far more than it ever did in a keyword auction. If your creative pipeline is thin, fix that before you move the budget.
Do not kill search entirely. Brand terms and your handful of proven bottom-funnel queries still convert. Shrink search down to what still works instead of switching it off.
And micro conversions are a bridge, not a destination. The end state is offline conversion tracking that sends qualified CRM stages back to the platform, the setup I detailed in our Salesforce to Google Ads offline conversion guide. Micro conversions keep the machine learning while volume is low. As qualified volume comes back, shift the weight to the real thing.
Search is not coming back to what it was. The accounts that get through this are the ones that give the platforms a new signal to learn from instead of waiting for the old one to recover.
If you would rather not guess at which signals to feed the machine, this is work we do every week. Request a free system review and we will look at your account together.
Sources
About the author
Kyle Rutledge
Owner
I’m Kyle, founder of Gradari, a paid ads lead generation agency that helps B2B and SaaS companies stop wasting budget on low-quality leads and start building systems that actually drive growth.