A campaign delivers the right searches, from the right places, at a sane cost per click. Conversions do not move. The meeting that follows is almost always about the ads.
It is usually about the page.
The distinction matters because the two have opposite fixes. An ad problem is solved by changing who arrives. A page problem is solved by changing what happens when they do — and no amount of bidding will fix it, because every improvement you make upstream gets poured into the same leaking container.
Here is how to tell which one you have, before anyone touches the account.
1. Compare conversion rate by query, not by campaign
Pull your top twenty converting-intent queries and look at conversion rate per query rather than per campaign. Campaign-level numbers blend traffic that has nothing in common.
If high-intent queries — the ones that name your product and a buying word — convert at roughly the same poor rate as everything else, the page is failing people who had already decided. That is not an audience problem. Someone searching “buy [your product] online” has told you exactly what they want.
If conversion rate drops off sharply as intent weakens, the page is doing its job and the campaign is buying too broadly.
2. Check whether people are reaching the form at all
Scroll depth and form-field focus events answer this in an afternoon.
- Visitors leave before the first screen finishes rendering → speed problem.
- They scroll, read, and leave without touching the form → the offer is not clear or not credible.
- They start the form and abandon it → the form is asking too much, too early.
Three very different fixes, and the account-level conversion rate alone cannot distinguish them.
3. Load the page the way the visitor does
Not on your laptop, on your office wifi, with the page already cached. On a mid-range phone, on mobile data, cold.
Most pages that “feel fast” to the team who built them take six or seven seconds in that condition. A measurable share of the people you paid for are gone before anything renders, and they will never appear in your bounce analysis as a page problem because they never became a session.
4. Read the first screen as a stranger
Cover everything below the fold and ask three questions:
- What is being offered?
- What does it cost, or what determines the cost?
- What happens after I click the button?
If a page cannot answer all three above the fold, visitors are being asked to commit before they have been told what they are committing to. That shows up in the account as a campaign that does not convert.
The tell that settles it
Run the same campaign to a different page — even a crude one, built in an afternoon, that answers those three questions plainly.
If conversion rate moves, you had a page problem. If it does not, the campaign is buying the wrong people and no amount of design will rescue it.
That test costs a day. Rewriting ad copy against a broken page costs a quarter and teaches you nothing.