For most small service businesses, keep control of the destination

If you run a local service business, professional practice, clinic, or any company where a click should produce a call, booked appointment, or qualified enquiry, do not let Google choose from your entire website by default. Start with the page you built for that search and treat Final URL expansion as a controlled test, not a permanent setting.

Google’s system is designed to find a page it considers relevant to the query. That is useful when your site has a clean catalogue of products or services. It is much less useful when the same domain contains a homepage, blog posts, FAQs, team biographies, booking confirmations, old offers, and pages written for existing customers. Relevance to Google is not the same as readiness to convert.

This is the same discipline behind optimizing advertising for booked appointments, not the cheapest leads. The landing page is part of the offer. If the page changes, the message, proof, form, and next step can change with it.

Choose the setting by the shape of your website

What Google’s automation actually changes

Google uses the term Final URL expansion for a setting that can replace the URL you entered with another relevant URL on the same domain. In AI Max for Search campaigns, the feature is enabled by default when you opt into AI Max. Google may also generate text that matches the selected page. In Performance Max, Final URL expansion is turned on by default, and Google recommends leaving it on while excluding pages such as blogs, careers, or FAQs that are not built to convert. (support.google.com)

That distinction matters. You are not merely giving Google permission to test a different page. You are allowing it to change the connection between query, ad copy, and destination. With AI Max, Google’s reporting now shows search terms, landing pages, and expanded URL assets together. The company specifically tells advertisers to use those reports to find weak landing pages and add exclusions. (support.google.com)

The practical implication is straightforward: expansion is only as safe as the pages Google is allowed to discover. A tidy ecommerce site with product pages, useful category pages, and accurate conversion tracking gives the system sensible choices. A small-business website assembled over several years often does not.

The numbers support testing, not blind trust

The practitioner pattern is more useful than the headline result

The positive case is real, but it is narrower than Google’s sales language suggests. Mudd Advertising reported a 190% conversion lift for one dealership and 153% for another after consolidating campaigns, tightening ad groups, waiting through a learning period, and deploying AI Max with URL expansion and text customization. Those results are encouraging for businesses with many related products or services, but the test changed the account structure at the same time. It cannot tell us how much came from the landing-page setting alone. (mudd.com)

The negative cases show where control matters. Budai Media found almost $110,000 in Performance Max spend going to a page with 0.95x ROAS that the client had not selected. Excell PPC describes a separate test in which AI Max sent paid traffic to a thank-you page, causing the page load to register conversions that had already happened. These are agency reports, not controlled industry studies, but they identify a failure mode that is easy to miss when you look only at campaign-level conversion totals. (thebudaimedia.com)

A smaller education account offers the more practical middle ground. Socio360 reports turning URL expansion off for Performance Max, mapping asset groups to three programs, and reducing CPA from $261 to $110 in seven days. That change happened alongside other account improvements, so it is not a clean experiment. Still, it reflects the operating choice I would make for a service business: give each offer a deliberate destination, then let automation work within a defined structure. (socio360.com)

A safer way to test Google Ads final URL expansion

1. Audit the pages Google can see

   Remove or noindex obsolete offers, thin pages, confirmation pages, private customer areas, and anything that should never receive paid traffic.

2. Build the control campaign first

   Use one deliberate final URL for each search theme and confirm that the form, call tracking, thank-you flow, and CRM attribution work.

3. Turn on expansion in a contained test

   Use a separate campaign, a limited budget, or a clearly defined group of queries so the test cannot quietly replace your main acquisition path.

4. Use inclusions where available

   A whitelist of approved pages is safer than trying to remember every page that should be excluded. Google’s AI Max reporting supports landing-page review and URL exclusions or inclusions. ([support.google.com](https://support.google.com/google-ads/answer/16470459?hl=en))

5. Judge booked business, not form volume

   Compare qualified calls, appointments, sales, and revenue by landing page. A cheaper lead from the wrong page is not an improvement.

6. Keep the winner and document the reason

   If expansion wins, record which pages and queries created the gain. If it loses, turn it off without rebuilding the whole campaign.

Give Google a better site before giving it more freedom

The best argument for automation is not that Google understands your business better than you do. It is that a well-structured website can give Google more useful options than one manually selected page can provide. That is a different claim, and it puts responsibility back on the site owner.

Create pages around real buying intents. Keep the offer, proof, pricing context, qualification, and next step together. Do not assume a generic services page can carry every ad group. Reuse the site’s structure where it helps, but keep the sales message specific, as explained in why a small business website should reuse its structure, not its sales message.

Then inspect the landing-page report as part of routine account management. Google says AI Max advertisers should wait at least two weeks after enabling the feature before making changes while the system learns. That is a learning window, not a reason to stop checking whether the traffic is going somewhere sensible. (support.google.com)

My recommendation as of September 11, 2026: keep Final URL expansion off for the core lead-generation campaign unless you have a specific reason to test it. Consider turning it on for ecommerce or a broad, well-organized service catalogue only after you have approved the page set, verified conversion tracking, and decided what outcome would justify keeping it. Let Google search for opportunities. Do not let it define the customer journey by default.