Google Knows Who Searched For Your Competitor - Here Is How To Reach Them

Someone typed your competitor’s name into Google this week. They read a couple of reviews, opened two websites, and closed both tabs. They haven’t decided yet.

You could chase them in Search. Bid on the competitor’s brand name, pay a high cost per click, and accept a poor quality score. Plenty of businesses do it, and in some markets it works.

There’s a second route that almost nobody uses. Instead of bidding on the search, you build an audience out of the people who made it. Then you reach them later on YouTube or in Gmail, at a fraction of the click price.

That’s what a custom segment does. It turns a search term into an audience.

Search Engine Land put numbers on the gap in its guide to custom segments. Terms costing more than twenty dollars a click in Search reached comparable users in Demand Gen. The Demand Gen price was closer to a dollar.

Treat that as an illustration rather than a promise. Your market sets your prices. But the direction of the gap holds up, and for a business on a small monthly budget, the direction is the whole point.

A custom segment is the closest thing Google Ads has to buying search intent without paying search prices.

Two things have to be true first. You have to know what the feature actually does, and you’ve to turn off the setting that undoes it. Most guides cover neither, so we will start there.

WHAT A CUSTOM SEGMENT ACTUALLY IS

A custom segment is an audience you define yourself.

Google gives you a long list of ready-made audiences. People in the market for a boiler. People with an interest in home improvement. Useful, and often too broad to be worth the money.

A custom segment lets you build your own instead. You supply the inputs, and Google finds people whose recent behaviour matches them.

You’ll still see the old names in older articles. Custom affinity and custom intent used to be two separate features. Google merged them into one thing called custom segments. If a guide tells you to choose between them, it’s out of date.

Here is the part that matters. A custom segment isn’t a remarketing list. Nobody in it has to have visited your website. You are describing a kind of person, and Google is going out to find them.

That makes it a prospecting tool, not a re-engagement tool. Judge it accordingly.

THE FOUR INPUTS, AND WHAT EACH ONE REALLY DOES

When you build a segment, Google offers four ways to describe the person you want.

Most advertisers tick all four, assume they mean roughly the same thing, and never look again. They do not mean the same thing, and one of them behaves nothing as it reads.

Interests, entered as keywords

You enter words and phrases that describe your ideal customer’s interests. Google reaches people likely to be interested in those topics, based on their behaviour.

This is the broadest input. It is closest to an affinity audience that you wrote yourself.

Search terms, entered as keywords

This is the powerful one. Google describes it as people who searched for any of these terms on Google properties.

That’s genuine search intent, taken out of the Search auction and made available as an audience. It is also where the competitor angle lives, because a competitor’s brand name is a search term like any other.

There’s a catch, and it’s important. This only behaves as searched-for on properties Google owns, such as YouTube, Discover and Gmail.

Run the same segment on the wider Display network, and Google demotes your search terms to interests. It doesn’t know those users the same way.

So the rule is simple. If you built the segment around search terms, run it where Google owns the inventory.

Website addresses

Here is the myth that costs people money.

Enter a competitor’s website, and it feels like you’ve just bought their audience. You haven’t.

Google’s own documentation is blunt. Entering a URL doesn’t mean your ads will show on those URLs. Your ads reach people who browse websites similar to the ones you enter.

Read that twice. It isn’t their visitors. It is people whose browsing looks like the kind of person who visits sites of that type.

That’s still useful. A trade supplier’s website, three industry magazines and two forums will describe a tradesperson quite well. Just do not build a plan on the belief that you are picking off a named competitor’s traffic.

If you genuinely want your ad to appear on a specific website, that’s a different feature. It is called placement targeting, and it lives in Display campaigns.

One more practical note. Use top-level domains rather than deep pages. Google is reading the site as a signal, not indexing that one article.

App names

Apps work exactly like websites, with the same caveat. Google reaches people who use apps similar to the ones you name, not the users of that specific app.

Useful in a few niches. Ignorable in most.

WHERE YOU CAN USE CUSTOM SEGMENTS, AND WHERE YOU CANNOT

This trips up more people than anything else on the list.

Custom segments work in Display, Gmail, Demand Gen and Video campaigns.

They do not work in Search campaigns. They do not work in Shopping campaigns.

That surprises people, because search terms are the strongest input and Search is the campaign everyone knows. You are allowed to build an audience from what people search. You aren’t allowed to point it at the Search results page.

Performance Max sits in the middle. You can’t target a custom segment there in the way you can in Demand Gen. You can feed one in as an audience signal, which tells Google’s automation where to start looking rather than where to stay.

That difference matters. A signal is a suggestion. Targeting is a boundary.

Search campaigns can still use other audience types, and Google recommends adding them on the Observation setting rather than Targeting. More on that in a moment, because the two settings do very different jobs.

CUSTOM SEGMENTS AGAINST IN-MARKET AND AFFINITY

Three audience types, and a straightforward way to choose between them.

Affinity segments group people by what they’re into over the long term. Their hobbies, their habits, the lifestyle they have settled into. Broad reach, low cost per thousand impressions, and a long way from a purchase.

Use affinity when the job is to be known. Brand campaigns, top of funnel, video that introduces you to people who have never heard your name.

In-market segments group people by recent buying signals in a category. Someone comparing washing machines this fortnight sits in the in-market segment for washing machines.

Use in-market when the job is to convert. It is the closest ready-made audience to a person with their wallet out.

Custom segments are for when neither list fits your business.

That’s the honest test. Google’s categories are built for volume, so they work well for mainstream products and badly for narrow ones. If you sell industrial pressure washers to contract cleaners, no ready-made segment describes your buyer. If you sell trainers, several do.

Build a custom segment when Google’s own categories are too broad to describe your customer, not because building one feels more sophisticated.

There’s also a quiet commercial advantage. A custom segment exists only in your account. Nobody else has built the same audience. You aren’t queueing up against every other advertiser for an identical, heavily bid list.

THE SETTING THAT QUIETLY UNDOES YOUR WORK

You can do everything above correctly and still end up advertising to people you never chose.

Optimised targeting is switched on by default on all campaigns. Its job is to find extra conversions. It goes after people it judges likely to convert, using live data from your campaign.

Read that again with your careful segment in mind. You have just spent an hour describing exactly who you want. Optimised targeting is designed to look outside that description.

Sometimes that’s welcome. If you’ve strong conversion data and you want scale, letting Google widen the net can work.

It isn’t welcome when you are testing. If you want to know whether your segment works, you can’t let Google add people to it while you measure. You’ll be reading a result that came from two different audiences and calling it one.

You turn it off in the ad group settings.

Turn optimised targeting off for the first month of any new custom segment. Judge the segment, then decide whether to let Google widen it.

This is the single most common reason a custom segment appears to do nothing, or appears to do brilliantly for reasons nobody can explain afterwards.

TARGETING OR OBSERVATION, AND WHY IT MATTERS

Every audience you add carries one more setting, and it changes everything the audience does.

Targeting narrows your reach. Add an audience on Targeting, and your ads only show to people in it. That’s what you want when the audience is the whole point, such as a Demand Gen campaign built on a search-term segment.

Observation doesn’t narrow anything. Your ads still show to everyone, and you simply collect data on how that audience performs against the rest.

Google recommends Observation as the default for Search campaigns, and for Display campaigns run by more experienced advertisers.

There’s a bonus with Smart Bidding. First-party audiences added on Observation feed your bidding strategy as signals. The campaign learns from them without being fenced in by them.

The mistake to avoid is adding a narrow audience on Targeting to a campaign that was already working. Traffic collapses overnight, and the audience takes the blame for a setting error.

WHAT MAKES A SEARCH TERM WORTH PUTTING IN A SEGMENT

The terms you choose decide whether this works. Everything else is admin.

Most people load the segment with category words. Roofing. Bookkeeping. Waste collection. Those words describe an entire industry, and the people typing them include students, journalists, job hunters and your competitors.

A good term has a buyer standing behind it. Here are the four tests I put every term through.

Does it imply a decision, not a topic?

“How does a flat roof work” is a person reading. “Flat roof replacement quote” is a person buying.

Both contain your service. Only one of them is worth paying to reach twice.

Would a person outside your market ever type it?

This is the quickest filter of the four. Say the term out loud and ask who else might search it.

“Commercial waste collection contract” has almost no innocent explanation. “Recycling” has a thousand.

Is it specific enough to be scarce?

Broad terms create huge audiences, and huge audiences behave like no targeting at all. If your segment could describe two million people, you haven’t narrowed anything.

Narrower terms build smaller audiences that actually mean something. You can always widen later.

Did it produce money in your own account?

This is the strongest test, and it costs nothing. Terms that already generated enquiries in your Search campaigns are proven buyers, in your market, at your prices.

Guessed terms are guesses. Proven terms are evidence. Start with the evidence and add guesses later, once you know what the baseline looks like.

Competitor brand names deserve a note of their own. A distinctive name is a superb term, because nobody types it by accident.

A descriptive company name is a poor one, because you can’t separate the brand from the category. Take a competitor called “Northern Drainage Solutions”. Searches for that phrase include people who just want drainage in the north.

HOW TO BUILD YOUR FIRST CUSTOM SEGMENT IN TWENTY MINUTES

Enough theory. Here is the build, start to finish.

Open Audience Manager. In Google Ads, go to Tools, then Shared Library, then Audience Manager, then Custom Segments.
Click the plus button and name it properly. Include the campaign type, the theme and the date, for example, “DemandGen – Competitor Brand Searches – Aug 26”. You’ll thank yourself in six months.
Choose the search terms option. This is the one described as people who searched for these terms on Google. It is where the intent lives.
Pull your terms from your own account. Open the search terms report on your Search campaigns and take the 20 to 30 terms that actually produced enquiries. These are proven, not guessed.
Add competitor brand names as separate terms. Distinctive names only. A competitor called Pinnacle Roofing Systems is a useful term. A competitor called Roofing Services London isn’t, because that phrase describes an entire market.
Leave the URL and app fields alone for this first build. Mixing inputs makes the result impossible to read. Get one signal working before you add a second.
Save it, then attach it to a Demand Gen campaign in the Targeting settings.
Turn off optimised targeting in the ad group settings before you launch. This is the step everyone skips.
Turn off Display network and video partners if your segment uses search terms. That keeps you on properties Google owns, where your terms behave as searches.

That’s the whole build. The research is the slow part, and most of it’s reading a report you already have.

Two things not to do. Do not build fifteen segments in one afternoon, because you’ll never be able to tell which one worked. And do not use your own brand terms. Those people already know you, and Search picks them up more cheaply.

A WORKED EXAMPLE, START TO FINISH

Numbers make this easier to follow, so here is a hypothetical business to work through. A commercial cleaning firm, eight hundred pounds a month, currently all of it in Search.

Their Search campaign works, in the sense that it produces enquiries. It also eats the whole budget by the third week of every month. Contract cleaning terms are expensive, and three national firms bid on all of them.

Step one, they open the search terms report for the last ninety days. They read it the other way round. Not hunting negatives this time, but hunting the terms that produced enquiries.

Twenty-two terms did. Things like “office cleaning contract quote”, “commercial cleaning company near me” and “end of tenancy cleaning for landlords”.

Step two: they add four competitor names. Two are distinctive enough to use. The other two are variations on “city cleaning services”, so they’re dropped for being indistinguishable from the category.

Step three: they build one custom segment from those twenty-four terms, using the search terms option only. No URLs, no apps. One signal, so the result can be read.

Step four: they build a Demand Gen campaign at fifteen pounds a day. That budget comes out of Search, not on top of it. Optimised targeting off. Display network and video partners off, so their terms behave as searches rather than interests.

Step five: they leave it alone for a month. This is the hard part.

What should they expect? Cheap clicks, almost certainly. Demand Gen will make their Search cost per click look outrageous by comparison.

They should ignore that entirely. One number matters. The cost of a qualified enquiry, set against the ninety pounds Search charges them.

They should also expect some rubbish. Prospecting reaches people earlier, so a share of the response will be domestic enquiries when they only do commercial. That’s normal, and it’s why the enquiries get counted by hand in month one.

If a qualified enquiry comes in under the Search figure, they have found a second channel, and they can shift more budget. If it comes in at triple, the segment was too broad, and the terms need tightening.

Either way, they learn something for four hundred and fifty pounds. That’s a fortnight of their existing Search campaign.

WHAT HAPPENS AFTER SOMEONE SEES THE AD

A custom segment gets you in front of the right person. It doesn’t close anything, and expecting it to is the fastest route to switching it off too early.

Almost nobody watches a video advert from a company they have never heard of and rings up that afternoon. The realistic outcome is that they now recognise your name.

That recognition is only worth money if you’ve a second step. The people who watched become an audience of their own, and that audience is far warmer than the segment you started with.

So the sequence runs in two stages. The custom segment finds strangers with the right intent. A retargeting audience built from the people who engaged brings the interested ones back.

Skip the second stage, and you are paying to introduce yourself over and over, then letting each introduction go cold. That’s the most common way a decent audience strategy still loses money.

Set the second stage up before you launch the first. It takes ten minutes, and it’s the difference between an advert and a funnel.

HOW TO TELL WHETHER IT IS ACTUALLY WORKING

Custom segments attract a particular kind of self-deception. The metrics look wonderful, and the bank balance doesn’t move.

Cheap clicks are the trap. Demand Gen and Video will hand you a cost per click that makes your Search campaign look extravagant. That isn’t a result. It is a different auction with different competition.

Judge a custom segment on three things.

Qualified enquiries, not conversions

Count the enquiries that were genuinely worth having. A form fill from a job hunter isn’t a conversion. Nor is a price request for something you do not sell. Nor is a wrong number.

This matters more here than anywhere else in your account. Prospecting audiences reach people earlier in their thinking, so a slice of the response will be curious rather than commercial.

Cost per qualified enquiry against your Search campaigns

This is the comparison that decides it. If Search delivers a qualified enquiry at ninety pounds and your custom segment delivers one at sixty, you’ve found something.

If it produces twenty conversions at four pounds and none of them answers the phone, you’ve found nothing. The dashboard will still look wonderful.

Assisted conversions, over a longer window

Prospecting audiences rarely close on the click. Someone sees your video, does nothing, and searches your name a fortnight later.

Look at your conversion paths before you switch it off. A segment that never converts directly but keeps appearing before your branded searches is doing exactly the job you hired it for.

Give it a month before you judge, and keep the settings still while you do. Changing the audience, the budget and the creative in the same week guarantees you learn nothing.

THE MISTAKES THAT WASTE THE MOST MONEY

A short list, in the order I see them most often.

Leaving optimised targeting on during a test. You end up measuring Google’s audience, not yours.
Believing the URL field targets that website’s visitors. It targets people who browse similar sites, which is a different and much larger group.
Using a search-term segment on the wider Display network. Your carefully chosen searches get demoted to interests.
Building the segment from generic category words. “Roofing” describes students, journalists and competitors as well as customers.
Adding a narrow segment on Targeting to a campaign that was already performing, then wondering where the traffic went.
Judging it on cost per click. Cheap traffic from the wrong people is the most expensive thing in the account.
Creating so many segments that no single one gets enough budget to prove anything.

FREQUENTLY ASKED QUESTIONS

What is a custom segment in Google Ads?

A custom segment is an audience you build yourself by entering keywords, search terms, website addresses or app names. Google then finds people whose recent behaviour matches those inputs.

You can use custom segments in Display, Gmail, Demand Gen and Video campaigns. You can’t use them in Search or Shopping.

Can I use custom segments in a Search campaign?

No. Custom segments aren’t available in Search or Shopping campaigns.

Search campaigns can still use other audience types, such as in-market segments and your own remarketing lists. Google recommends adding those in the Observation setting, so they inform your bidding without cutting your reach.

Does entering a competitor’s website URL show my ads on their site?

No, and this is the most expensive misunderstanding in the whole feature.

Google states that entering a URL doesn’t mean your ads will show on those URLs. Your ads reach people who browse websites similar to the ones you enter. If you want your ads on specific sites, use placement targeting in a Display campaign.

What is the difference between a custom segment, an in-market audience and an affinity audience?

Affinity segments group people by long-term interests, which suits awareness work. In-market segments group people showing recent purchase intent, which converts best.

Custom segments are the ones you define yourself, for when Google’s ready-made categories are too broad to describe your buyer.

Why is my custom segment getting no impressions?

Three usual causes, in the order to check them.

The audience is too small because the inputs were too narrow. The campaign type doesn’t support custom segments, which rules out Search and Shopping. Or the daily budget is too low to enter the auction consistently.

Widen the inputs first. That fixes it most of the time.

How many keywords should I put in a custom segment?

Around 20 to 30 search terms is a sensible starting point. Below roughly ten, the audience is often too small to serve.

The quality of the terms beats the quantity every time. Use phrases that show someone choosing a supplier, not someone reading up on a subject.

Do custom segments work on a small budget?

They can work better on a small budget than Search does, for one specific reason. You are the only advertiser with your exact segment, so you aren’t bidding against everyone else for an identical list.

Costs per click on Demand Gen and Video are usually far lower than Search. That gives a modest budget more chances to be seen, as long as you measure qualified enquiries rather than clicks.

WHERE TO START THIS WEEK

You do not need a new campaign to get value from this.

Open your search terms report and find the terms that produced real enquiries over the last ninety days. That list is the most valuable targeting asset you own, and most businesses use it once for negatives and never again.

Build one custom segment from those terms. Add three or four competitor brand names if they’re distinctive enough to be worth having. Attach it to a single Demand Gen campaign, turn off optimised targeting, and leave it alone for a month.

One segment, one campaign, one honest measurement. That’s a real test, and it costs less than most businesses spend on Search in a week.

The businesses that win at this aren’t the ones with the cleverest audience. They are the ones who checked what the setting actually does before they trusted it with their budget.

Did You Enjoy This Blog Post?

I hope you enjoyed this blog post, and thank you so much for being here. We also upload videos to our YouTube channel every weekday. Please subscribe so you are one of the first to be notified.

If you enjoyed this blog, you may also like: