Signs that my search engine ads are driving low-quality traffic?

Why search ads hide low quality better than display

Search traffic looks honest on paper. The user typed something. They clicked an ad that matched the query. The session counter goes up. None of that tells you whether the click had any purchase intent behind it, and on Search the noise is harder to see because every visit feels earned.

The six patterns below sit in reports a founder rarely opens. Each one has a threshold worth memorizing and a fix that lives in the campaign, not in the landing page.

Pattern 1: high session counts with page depth of one

Open GA4. Filter to Google Ads source. Pull the engagement report and look at pages per session. If the median sits at 1.0 or 1.1 across a campaign that has run for thirty days, the traffic is bouncing on arrival without scrolling, clicking, or triggering a second pageview.

The threshold to watch is page depth under 1.2 on more than sixty percent of sessions. That is the line where the traffic stops behaving like buyers and starts behaving like accidental clicks.

The fix is rarely the landing page. It is the match type. Phrase and broad match on a tight budget pull queries the bidder thinks are close enough, and “close enough” is a long way from “ready to buy.” Tighten to phrase plus a negative list, or move to exact for the spine keywords.

Pattern 2: qualified-looking queries converting at near zero

The search-term report is the most honest document in a Google Ads account. Filter to the last ninety days, sort by cost descending, and look at the conversion rate column next to queries that read like buyer intent.

If a query like “[product name] price” or “[service] near me” is pulling fifty clicks and zero conversions, the query is not the problem. The landing experience after the click is breaking the promise the ad made. Read the headline of the page that received the click. Read the headline of the ad. If they do not say the same thing in the first three seconds, the visitor leaves.

The threshold here is any commercial-intent query with twenty-plus clicks and a conversion rate under 0.3 percent. That is the trigger to pull the ad-to-landing-page pair and reread it as the buyer, not the marketer.

Pattern 3: mobile traffic share above eighty percent on a desktop offer

Pull the device report. If the offer requires a desktop interaction such as a long form, a high-consideration purchase, or a configurator, and mobile sits above eighty percent of clicks, the campaign is being served to the wrong context.

This happens because mobile clicks are cheaper, the bidder optimizes for click volume by default, and the platform never asks whether the offer can close on a phone. Set a device bid modifier of negative forty percent on mobile for desktop-converting offers, or split desktop and mobile into separate campaigns with separate landing pages.

The device-split decision lives in the campaign-structure section of the setup workbook referenced below.

Pattern 4: dwell time under ten seconds across a campaign

Average engaged session duration under ten seconds means the visitor did not read the page. They saw it, scanned the top fold, and left. Ten seconds is shorter than the time it takes to scroll past the hero on most sites.

Cross-reference this with Core Web Vitals before changing targeting. If the Largest Contentful Paint is over four seconds, the page is losing buyers to load time, not to relevance. Fix the page first. If LCP is healthy and dwell is still under ten seconds, the ad copy is overpromising what the page delivers.

Pattern 5: repeat-visitor spam on the auction insights report

A pattern that hides on Search but ruins budgets: the same IP ranges, the same competitor domains, or the same data-center signatures showing up repeatedly in click-fraud filters and in the auction insights report.

Open auction insights filtered to the last thirty days. If a single competitor shows above sixty percent overlap rate on your branded terms or on a non-branded keyword group, something is bidding against you with abnormal consistency. Combine that with a sudden invalid-click filter rate above fifteen percent in the billing report and the traffic is being inflated by something that is not a human buyer.

Google refunds invalid clicks automatically, but only the ones it catches. The ones it misses sit in the conversion model and skew bidding. Adding IP exclusions and tightening location targeting to presence-only is the structural fix.

Pattern 6: search-term-to-landing-page topical mismatch above twenty percent

This is the pattern that catches the most spend in audits. Pull a sample of fifty top-spending search terms. For each one, open the landing page that received the click. Read the H1.

If twenty percent or more of the pairings have a topical gap (the query is about one thing, the H1 is about another), the campaign is buying clicks that the page cannot close. Final URL expansion on Performance Max and dynamic search ads make this worse by default, because the platform picks the landing page on its own.

Turn off final URL expansion. Map every ad group to one landing page with one promise. If a query needs a different page, it needs a different ad group.

The Tracking Stack reference covers the event setup that makes this audit reliable. Without conversion events tied to scroll depth and engaged session, dwell and depth numbers are guesses.

Fixing the leak structurally

Three structural moves catch most of the patterns above. Tighten match types and final URL expansion at the campaign level. Split desktop and mobile when the offer demands it. Map every ad group to one landing page with one promise.

The diagnosis library at /wasted-ad-spend/ walks through the rest. The Google Ads Setup Audit is the working document a founder can run against a live account in an afternoon. Both are free, no email gate, and worth reading before the next budget review.

Founders who want a senior pair of eyes on the low-quality-traffic diagnosis instead of running it solo can book a thirty-minute call.