How to monitor your brand terms in paid search

Most brand owners find out about a problem by accident — a customer mentions a confusing ad, or someone on the team happens to search the brand name and spots something odd. By then the infringer has often been running for weeks. The auction for your brand term doesn’t hold still: new advertisers enter, existing ones test new copy, and affiliates rotate in and out, often on a schedule that has nothing to do with when you happen to look.

That’s the core problem with treating this as a search you run when something feels off. A rising cost-per-click is one sign it’s time to look — see our guide on why your brand CPC is rising — but waiting for a symptom means you’re always finding out after the fact. Real monitoring means having a standing answer to “who’s on my brand term right now,” not reconstructing it after a bill or a customer complaint prompts you to check.

The three tools below each answer a different slice of that question. None of them is wrong to use — the mistake is picking one and assuming it covers the other two.

Auction Insights, on your brand campaign in Google Ads, is the natural starting point because it’s already there and it’s free. For any set of keywords it shows six metrics per competing domain: impression share, overlap rate, position above rate, top-of-page rate, absolute top-of-page rate, and outranking share. Read across a few weeks, it tells you who’s regularly present on your brand term and whether their presence is growing.

• It’s aggregate and historical — you get rate statistics for a domain, not the actual headline, description, or landing page of any specific ad, and not a live view of what’s running right now.
• It goes dark exactly when you need it most: Google withholds the data once your own impression share on those terms drops below 10%, which is often the early signal of a new, aggressive entrant.
• It’s UI-only. There’s no Auction Insights endpoint in the Google Ads API, so you can’t pull it into a dashboard or script an alert on top of it — someone has to open the report and read it.

There’s also a distortion worth knowing about specifically because it’s recent. In April 2025, Google changed its Unfair Advantage policy to let the same advertiser run more than one ad on the same results page, each in a different ad location — for example, one at the top and one lower down. That was previously prohibited. The practical effect: a competitor’s impression share in Auction Insights can climb because they’re now serving two slots on one search rather than because they’re bidding harder, and the report has no way to tell you which. Impression share on its own is a weaker signal now than it used to be. Look at it over a 30–90 day window and read it alongside overlap rate and position above rate together, rather than reacting to a single spike.

Searching your brand term yourself is the one method that shows you the truth: the actual ad, the actual headline, the actual landing page, exactly as a customer would see it. Run it in an incognito window set to your target location, or use the Ad Preview and Diagnosis tool so repeated checks don’t inflate anyone’s impression counts, including your own. When you need to know precisely what an ad says — for a trademark complaint, for instance — this is the only method that gives you evidence you can act on.

The problem is what it takes to do this properly at any real scale. A business rarely has one brand term; it has variants, misspellings, product-line names, and the brand plus category modifiers. Multiply that by every market you operate in, by desktop and mobile, and by the fact that infringers rarely run all day — they surface at particular hours or in particular regions, so a single morning check from one location can miss them entirely. Covering that properly by hand means dozens of searches, repeated on a schedule, across time zones — which nobody does consistently for more than a few weeks before it quietly stops happening.

Manual searching stays essential as a verification step. It’s just not a monitoring system on its own — it’s what you reach for once something else has told you where to look.

Automated, continuous monitoring isn’t a replacement for the other two — it’s what closes the gap between them. It runs the equivalent of the manual search, on your actual terms, across your actual markets and devices, on a schedule no person is going to sustain by hand, and it captures what it finds as timestamped evidence rather than a fading impression of what you saw once.

• It catches intermittent and regional infringers — the ones who only run at certain hours or only target certain locations — because it isn’t limited to when someone remembers to check.
• It gives you dated, screenshot-level evidence of the actual ad the moment it appears, which is exactly what a trademark complaint or an internal escalation needs.
• It alerts you on day one. The alternative — finding out in a monthly report, or from a customer, or from a cost spike you eventually trace back — means the infringer has usually had a run of several weeks before anyone acted.

The output is the same kind of evidence manual searching gives you — an actual ad, actually seen — just gathered on a schedule and at a scale a person can’t maintain, and stored so you have a record instead of a memory of what you saw.

Put together, the three methods aren’t competing choices — they’re different resolutions of the same picture, and a sensible routine uses each for what it’s actually good at.

• Weekly or biweekly: skim Auction Insights on your brand campaign for new domains or a meaningful shift in overlap rate or position above rate, reading trends over weeks rather than reacting to one day.
• As needed: when Auction Insights flags something, or a cost or conversion number looks off, run a manual search from the right location to see the actual ad and confirm what you’re dealing with before you act on it.
• Continuously: run automated monitoring underneath both, so the gaps neither one covers — the advertiser who only appears at 2am in one region, the one who shows up and disappears in a single afternoon — don’t depend on someone happening to look at the right moment.

The honest version of this is that Auction Insights and manual searching are things you do; continuous monitoring is something that runs whether or not you remember to check. If the last few weeks have been “I should really look into that,” that’s the tell that checking by hand has already stopped working.