Case Study: Raising Lead Quality With AI Max for Search in Financial Services
How a US financial advisor matching platform got fewer form fills from Google Search and far more of the investors it actually sells to.
81%
Of requests from $250K+ investors, up from 55%
60%
Of qualified requests reached, up from 45%
30 days
From a zero-revenue first week to break-even
The Client
A marketplace that only gets paid when an advisor accepts the lead.
The client runs a US platform that matches individual investors with vetted financial advisors. A lead earns money only when an advisor on the other side accepts it, and advisors want people with $250,000 or more to invest. A form fill from anyone else costs media and returns nothing.
Their Google Search account had run for years and was not broken. Leads converted at 4.5% of clicks, and the client reached about 45% of the people who filled the form.
The Challenge
A healthy Search account that had run out of people to find.
Search only reached people who typed something close to a keyword the account bid on. Plenty of qualified investors search around the edges of that, with questions about fees, a life change or an account they already hold, and none of it looked like a keyword.
Performance Max was the first attempt to reach them. Over six weeks it returned about 55 cents per dollar, and much of the spend went to YouTube and mobile placements rather than to people looking for an advisor.
0.55
Revenue per dollar from Performance Max over six weeks.
4.5%
Historic Search form fill per click, the bar to beat.
A mature Search programme setting a high bar for any new campaign.
Revenue tied to $250K+ investors, not to form volume.
Performance Max had already returned less than it spent.
A regulated category where a page cannot say whatever converts.
What our audit found
AI Max on defaults read our pages literally.
The first week of AI Max on Google's defaults earned nothing. More than half of its clicks came from people trying to log into a retirement account they already had: "401k login", "walmart 401k", "find my old 401k". Every $250K+ lead it produced was rejected.
Tracking was the first suspect, and GA4 showed it was working. The real problem was who the campaign had decided to talk to. Google had seen "401(k)" on the pages and matched everyone who searched it.
Over half of week-one clicks were 401(k) login searches.
Every $250K+ lead rejected.
Conversion tracking verified clean in GA4.
Product and life-event pages drawing the wrong investors.
The Solution
How we turned it around.
197 pages in seven days
We built 197 landing pages on six templates in seven days. Three of those days went on design and on what helps an undecided reader: checklists, slider calculators and cited sources. AI agents drafted and researched each page and tested every source, then people reviewed every page before launch, because this is someone's retirement money.
Each page answers one context a person searches from, with one ad group mapped to it. Final URL expansion stayed on, and Google has since chosen 148 of the 162 pages that received clicks, mostly state and life-event versions of the core pages.
What we shipped
- Six templates, advertorial format.
- One ad group per intent and page.
- Final URL expansion on, headline rewriting on.
- Human review of every page before launch.
A negative list, then more budget
A dedicated negative list took the login searches to almost nothing within days. Conversions per day rose about sevenfold.
About two weeks after launch the PPC lead tripled the daily budget. The reason was quality: the leads now coming in were getting approved, and the campaign was losing 76% to 88% of its eligible impressions to budget. The next day brought a run of approved $250K+ leads.
What we shipped
- Negative list built from the search terms report.
- Budget tripled on approval rate, not on volume.
- Impression share lost to budget used as the safety check.
Engagement picked the ad. Revenue picked the page.
Every page is reviewed weekly on revenue returned per dollar, not on click-through. The life-event page had one of the best click-through rates in the account and still returned 0.74, because its visitors needed a one-time asset split rather than an advisor. It was paused, along with the product pages.
Pages about choosing an advisor got about 70% of their qualified leads approved. Product and life-event pages managed one approval between them, on identical settings. We treat that as the strongest pattern we have, not a rule, because the samples are small.
What we shipped
- Weekly revenue per dollar by page.
- Paused pages: 27.8% of spend returning 0.38.
- Live pages returning 1.14, or 1.25 counting the parent page.
The Numbers
Outcomes we can talk about.
The form fill rate went down, from 4.5% of clicks on historic Search to 2.9% on AI Max. Everything that pays went up. 81% of requests now come from people with $250,000 or more to invest, against about 55% on Search, measured the same way. The client reaches 60% of qualified requests, up from about 45%.
Fewer people ask, and the people who ask are the ones advisors want. After a first week that earned nothing, the campaign broke even within its first month. In a sample Monday-to-Sunday week, the whole campaign returned 1.01 per dollar including the three pages we were already pausing, and the four live pages returned 1.14 on their own.
81%
Of requests from $250K+ investors, up from 55%
60%
Of qualified requests reached, up from 45%
30 days
From a zero-revenue first week to break-even
My note to the team after the first week was "Something is wrong." The fix was not more keywords. It was telling Google who not to send us.
- Head of Digital Marketing, Intelegencia
What We Built
What's Next
Phase 2: 400 pages and every context.
Four months in, the work has moved from proving the campaign to scaling it. Phase 2 takes the site past 400 landing pages and grows the campaign to cover every context the client serves, with the same weekly rule: a page stays live because of the revenue it returns, not the clicks it gets.
Frequently Asked Questions
About This Project
The questions teams usually ask when they want to run a similar engagement.
AI Max with final URL expansion sends people to the page closest to their question, and many read it without filling the form. The share of requests from $250K+ investors rose from about 55% to 81%, so fewer requests produced more revenue.
The Real Numbers
Need real numbers? Let's talk.
We kept the names off the page. The story is real, the outcomes are real, and we're always happy to walk a serious team through the rest of it.
Start a Conversation→More from Marketing
Reducing Cost Per Lead With ChatGPT Ads in Financial Services
Reducing ACoS and Scaling Sales for an Amazon Seller
Scaling B2B Lead Generation for an Electrical Equipment Exporter
Insights from our blog
Digital MarketingWhat to Actually Review in Your Quarterly Marketing Meeting (And What to Stop Wasting Time On)
Learn how to run effective quarterly marketing reviews by evaluating qualified leads, campaign ROI, channel performance, and revenue impact to make smarter marketing decisions.
Digital MarketingAI Max for Search: Why It Paid for Our Lead Gen Client When It Fails for Most
Our first week of AI Max earned nothing and matched people trying to log into their 401(k). By the end of the first month it had broken even. What changed, from a US financial advisor matching account.
Digital MarketingWhere to Actually Spend Your Digital Marketing Budget in 2027
Learn where to allocate your 2027 digital marketing budget, balance paid and organic channels, strengthen measurement, and prioritize investments that drive measurable business growth.