Final URL expansion on a service site: is Google picking the right page?

Final URL expansion sounds harmless. Google sends clicks to whatever page it thinks matches the query, and Smart Bidding will keep spending on those visits as long as they convert at some rate. The catch is that "whatever page" often turns out to be a blog post or a team bio. Say a dental clinic has a services page, a meet-the-team page and a post about whitening. A search for "emergency dentist" might land on the whitening post because it mentions emergencies, and the booking form there is buried. Conversions still trickle in, so the campaign looks fine, but a page nobody built for that intent is doing the selling. I'd switch URL expansion off and only turn it back on for a specific page group after watching the landing page report for a few weeks. Has anyone seen it work well on a site with lots of content pages?

I'd keep it on and fence it off. Campaign-level URL exclusions can block /blog/ and /team/ while still letting Google expand to your service pages, so "emergency dentist" stops landing on the whitening post. The catch is that exclusions match URL prefixes, so this only works if the content sits under clean paths. If the blog posts live at the root of the domain, you'd have to list each one, and switching expansion off might be less work.

I'd pull the landing page report before writing any exclusions. Sort the last 30 days of final URLs by cost and you'll see which pages are soaking up the spend, and whether the root-level posts are a handful or dozens. It also gives you a baseline. If the whitening post is taking emergency-intent clicks, its share of clicks should drop once the exclusion is live, and you can check whether the service pages pick up that traffic or it just disappears.

On the baseline, I'd go one step further and check the conversion action, not just the clicks. A whitening post can pull cheap clicks that count as "leads" because someone filled in a newsletter or call-tracking form, so the cost-by-page view looks fine while the real bookings come from somewhere else. Split the report by conversion action and you'll see whether those stray pages are producing actual appointments or just noise. And if the noisy action is the only one being counted, Smart Bidding will happily optimize toward it.

Demote the soft actions to secondary before you split anything. Newsletter signups and call-tracking events can stay in the account for reporting, but if they're marked primary, Smart Bidding treats them as the goal. Only actions that represent a real appointment should feed bidding.

For the URL question, I'd use a simple rule: if a page has cost about three times your target CPA and produced zero primary conversions, exclude it. Below that, you usually don't have enough clicks to call it, because a page can look dead after a few unlucky visits. Above that, the gap is unlikely to be noise, and waiting longer just pays for the same mismatch.

It depends on volume, of course. A clinic getting ten conversions a month needs a longer window than one getting a hundred. But at least you get a number to argue with instead of a feeling.