Where Half Your Display Budget Quietly Goes

There’s a report in your Google Ads account that lists every website and app your Display ads appeared on, and what each one cost you.

Most advertisers have never opened it. The ones who do tend to go quiet for a minute.

You’ll find a flashlight app. A mobile game aimed at seven-year-olds. Three sites you’ve never heard of that publish nine hundred words of nothing across four pages, surrounded by adverts.

None of those people was looking for a commercial cleaner, or a solicitor, or whatever it is you sell. Several of them didn’t mean to click at all.

The waste isn’t hiding. It’s itemised, in a tab nobody opens.

And there’s now a deadline for doing something about it, which is new this year.

WHAT CHANGED IN MAY

Google announced in May 2026 that standalone Display campaigns are being absorbed into Demand Gen.

The timeline runs like this. A migration tool appeared in accounts from June 2026. From January 2027, you can’t create new standalone Display campaigns. Remaining ones get migrated automatically during 2027.

So Display isn’t dying exactly. It’s moving house, into a campaign type that behaves differently.

Why that matters for waste

Demand Gen is a more automated campaign type than Display. It decides where your ads run across YouTube, Discover, Gmail and the wider network. You steer it far less directly.

More automation is not automatically worse. It often performs well.

But the trade is control, and control is exactly what the placement problem needs.

So the sensible order is this. Audit now, while the reporting is granular and exclusions are easy to apply. Carry the lists into the migrated setup. Don’t wait until the controls thin out and then try to work out what went wrong.

Do the placement audit before you migrate, not after.

If somebody else runs your account, that’s the single question worth asking them this month.

WHERE THE MONEY ACTUALLY GOES

Three leaks account for most of it. They’re different problems, and they need different fixes.

Made-for-advertising sites

These are websites built to earn ad revenue rather than to be read. Thin content, enormous ad density, one article chopped across five pages so it can carry five times the adverts.

They exist for one reason. Somebody buys cheap traffic, points it at a page stuffed with ad slots, and pockets the difference.

Research by the ANA found 21 per cent of programmatic impressions landed on made-for-advertising sites. That’s an industry figure across the open market, not a measurement of your account. It still tells you the scale of the thing.

Your ad technically appeared. A human technically saw it. Neither fact is worth what you paid.

Mobile apps, and games in particular

This is usually the biggest single line in a small account’s placement report.

Ads inside mobile games sit beside a play button and a close button that moves. Add a child’s thumb. The clicks are real; they’re just accidental.

You pay for those clicks the same as any other. The visitor lands on your site, has no idea why, and leaves in two seconds.

Flashlight apps, battery savers and free games are the usual suspects. If you sell to businesses, essentially none of that inventory is worth buying.

The very long tail

Then there’s the tail. Hundreds of sites taking a pound or two each.

Individually invisible. Collectively, often the largest number on the page once you total it.

The tail is why “we only spend a bit on Display” isn’t reassuring. Small amounts across many bad placements still add up to a real number.

WHY THERE IS RUBBISH IN THE NETWORK AT ALL

It’s worth understanding why this happens, because it explains why the problem never fully goes away.

The Display Network isn’t a place. It’s an arrangement between millions of website owners and app developers who agreed to show adverts, and advertisers who agreed to pay for space.

Google sits in the middle and takes a cut. The more inventory in the pool, the more auctions it can run.

That’s a reasonable business. It also means the pool contains everything from a respected trade publication to a wallpaper app made in an afternoon.

The incentive problem

Whoever owns a placement gets paid when your ad appears on it. That’s true of the trade magazine, and it’s true of the puzzle game.

So there’s a standing incentive to create inventory that generates impressions cheaply, regardless of whether anyone useful sees them.

Made-for-advertising sites are that incentive followed to its conclusion. Buy traffic for less than the adverts pay out, keep the difference, repeat.

Nothing about it is illegal, and much of it isn’t even against the rules. It’s arbitrage, and it works because most advertisers never check.

Why automation doesn’t solve it for you

The reasonable response is that Google’s systems should filter this out, and to be fair, they do filter a great deal.

But the system is optimising towards your conversion goal using the signals it has. If it can buy a thousand cheap impressions in a game app and a handful of people accidentally click through, that can look acceptable in the data.

It doesn’t know that nobody in that app has ever bought commercial flooring. You do.

That’s the whole argument for spending twenty minutes on the placement report. You’re supplying the judgement the system doesn’t have.

HOW TO FIND YOUR OWN NUMBER

I’m not going to tell you what percentage of your budget is wasted, because I don’t know, and neither does anyone else writing about this.

Several pages on this topic quote 40 to 60 per cent without sourcing it. Ignore that. Your account has its own answer, and you can have it this afternoon.

Here’s the twenty-minute version.

Open the placement report for your Display campaigns and set the date range to the last 90 days. A short window will not show you the tail.
Sort by cost, highest first. Not impressions, not clicks. You are asking where the money went.
Go down the first fifty rows and mark each one: could a real customer plausibly have been here? Not “is this a real website”, but “would my buyer be on it”.
Add up the cost of everything you marked no. That is your number, and it is the only honest one you will get.
Now filter for app placements specifically and total those separately. In most small accounts, this is the single largest block.
Check the conversions column against the same rows. Cheap clicks with no conversions are the pattern you are looking for.

Two things people get wrong when they do this.

Don’t judge a placement on click-through rate. High CTR on a game app means accidental clicks, not interest. It’s a warning sign, not a good sign.

And don’t stop at the top twenty. The tail is where the quiet money is, which is the whole point.

THE EXCLUSIONS WORTH ADDING TODAY

Once you’ve seen your number, this part is quick.

Start with the app inventory

Excluding adsenseformobileapps.com removes app inventory broadly, and for most business-to-business advertisers that alone is the biggest single improvement available.

You can also exclude whole app categories, such as games or entertainment, if you’d rather keep some app inventory than none.

For individual apps, the syntax is a prefix followed by the app’s bundle ID. It looks like mobileapp::2- and then the identifier, for example mobileapp::2-com.example.somegame.

Build the list at account level, not campaign level

Account-level exclusion lists apply across campaigns. You do the work once, rather than every time somebody builds a new campaign.

You’ll find them under Tools, then Shared Library, then Placement exclusions.

The ceiling is generous. You can hold up to 65,000 placements across as many as 20 lists, which is far more than a small account will ever need. New exclusions usually take effect within about 12 hours.

Use content exclusions too

Content exclusions are separate from placements. They keep your ads away from categories of content rather than named sites.

Worth setting deliberately rather than leaving at the default, particularly for professional services where context matters to how you’re perceived.

One note on parked domains, because older guides still cover it at length. Parked domains stopped being an ad surface in the Search Partner Network in February 2026, and new accounts had already been opted out by default since October 2024. If an article is still telling you to exclude them manually, it’s out of date.

Don’t paste in a hundred thousand exclusions

The big downloadable lists are tempting, and I’d use them carefully.

They’re built from other people’s accounts. Some of what they block would have been fine for you. A very large list is also hard to audit later, when something stops working.

Start with your own placement report, which is evidence from your own money. Add a published list afterwards if you want the extra coverage.

TELLING A BAD PLACEMENT FROM A SLOW ONE

Not everything without a conversion is waste, and this is where people over-correct.

Cut too hard, and you strangle the campaign. Cut too little, and you keep funding games. Here are the four questions I’d ask of any row before excluding it.

Could my buyer plausibly be here?

This is the first filter, and it settles most rows on its own.

A trade magazine, a local news site, a forum about your industry. Those are places your customer might sit. A children’s puzzle game is not, whatever the numbers say.

If the answer is a clear no, exclude it and don’t look at the metrics at all.

Is the click rate suspiciously high?

Counter-intuitive, but a very high click-through rate on app inventory usually means accidental clicks rather than interest.

Fingers on small screens near a moving close button. That’s not engagement; it’s geometry.

Treat an unusually high rate on a game or utility app as a red flag rather than a discovery.

Has it had a fair chance?

A placement with eleven impressions and no conversions hasn’t told you anything yet.

Give a plausible placement enough volume to prove itself before you judge it. Excluding on tiny numbers is how people accidentally shrink a campaign to nothing.

Is it assisting rather than closing

Display rarely closes. Check whether a placement appears in conversion paths before deciding it did nothing.

A site that never converts directly but keeps turning up early in journeys that closed elsewhere is doing the job Display is actually for.

The rule I’d apply: exclude on relevance immediately, exclude on performance only with enough data to be sure.

WHAT ABOUT PERFORMANCE MAX

Worth a note, because the same waste turns up there and people assume nothing can be done.

Performance Max runs across the same inventory, including apps and the open network. For a long time, advertisers had very little visibility into where it went.

That has improved. Placement reporting exists, and placement exclusions can be applied, though the controls are less direct than Display gives you.

The practical advice is the same. Build your exclusions at account level so they apply as broadly as possible, rather than trying to manage each campaign separately.

If most of your budget sits in Performance Max, the app exclusion is the one to prioritise. It’s the single leak most likely to be quietly consuming budget with nothing to show.

We cover Performance Max reporting in more depth separately, because it hides more than placements from you.

A WORKED EXAMPLE, START TO FINISH

Here’s a hypothetical to make the process concrete. A commercial flooring company, six hundred pounds a month, split between Search and a small Display campaign.

The Display campaign spends about a hundred and eighty pounds a month. It shows plenty of clicks, a decent click-through rate and almost no enquiries. Nobody has questioned it in a year, because the click cost looks cheap.

They open the placement report, set it to 90 days, and sort by cost.

The top row is a mobile game. Forty-one pounds over the quarter, a click-through rate three times the account average, zero conversions and an average time on site of a few seconds.

Rows two through six are apps too. Two puzzle games, a torch app, a weather app and a wallpaper app. Another sixty-odd pounds between them.

Then a stretch of websites nobody recognises. Long domain names, articles about nothing, the sort of pages that split five hundred words across four screens. Another fifty pounds.

Further down, the useful stuff. A regional construction news site with two conversions. A facilities management blog with one. Together, about thirty pounds and three enquiries.

So the honest sum for the quarter: roughly a hundred and fifty pounds went to inventory their buyer would never be on, and about thirty pounds went to inventory that worked.

They exclude app inventory at account level, add the handful of made-for-advertising domains from their own report, and leave everything else running.

The next quarter the campaign spends less, reaches fewer people, and produces the same three enquiries at a fraction of the cost.

Nothing clever happened. They read a report and turned off the parts that were never going to work.

WHAT CARRIES OVER INTO DEMAND GEN

This is the practical question if you’re migrating, and the honest answer is that it varies by account.

Account-level exclusion lists are the ones most worth having in place, because they sit above individual campaigns rather than inside them.

So the order matters. Build the exclusions at account level first, then migrate. Doing it that way round means the protections are already sitting there when the campaign changes shape.

After migrating, check the placement reporting you get. If it’s thinner than you had before, that’s the trade you’ve made, and it’s another reason the audit was worth doing while the data was rich.

Then watch the first month closely. A migrated campaign is effectively a new campaign, and it will go through a learning period, whatever the interface calls it.

WHEN DISPLAY IS ACTUALLY WORTH IT

I’ve spent most of this post on what goes wrong, so let me be fair about what goes right.

Display works well for retargeting

This is the strongest case by a distance. Showing an ad to somebody who already visited your website is a different proposition from showing one to a stranger in a puzzle game.

They know who you are. The banner is a reminder rather than an introduction. Costs are low, and the audience is small and specific.

If you only ever use Display for retargeting, you’ll get most of the value with a fraction of the waste.

It works for reminding a warm audience

The viewer audience from a video campaign, your email list uploaded as customer data, people who abandoned a form. All of these are worth reaching cheaply, and Display does that well.

That’s the same funnel logic as our posts on custom segments and video ads. Find them somewhere, capture them, then remind them cheaply.

It’s weak as cold prospecting on a small budget

Using Display to find brand new customers is where most of the disappointment comes from. No audience, modest budget, cold inventory.

You’re buying attention from people doing something else. The inventory includes a lot of rubbish. And you’ve no relationship to fall back on.

Bigger budgets can make that work because they can afford the testing. A business spending a few hundred a month usually can’t.

And if your Search campaigns are still losing impression share on the terms that already make you money, that’s where the next pound belongs. Existing demand is always cheaper than manufactured demand.

WHAT GOOD LOOKS LIKE AFTER THE CLEANUP

Expectations matter here, because the numbers move in a direction that looks alarming if nobody warned you.

Your reach will fall. Sometimes sharply. You’ve just removed a large slice of available inventory, so fewer people see your ads.

Your click volume will fall too, and often by more than your spend does. That’s the accidental clicks disappearing, which is exactly what you wanted.

Your click-through rate may well drop. Games and utility apps generate a lot of clicks, so removing them takes the average down. A lower rate here is a sign of health, not decline.

What should hold steady, or improve, is enquiries. Same conversions on less money is the result you’re after. If enquiries fall in proportion to spend, you cut something useful and should look again.

Give it a fortnight before you judge. Exclusions take up to about 12 hours to take effect, and the campaign needs a little time to settle into the smaller pool.

One thing to watch for. If reach collapses to almost nothing, you’ve probably over-excluded, most likely by pasting in a very large downloaded list. Roll that back before you touch anything else, because it’s the most common cause.

And keep the habit. A quarterly look at the placement report catches new rubbish as it appears, and it takes ten minutes once the big offenders are already blocked.

THE MISTAKES THAT WASTE THE MOST MONEY

Never opening the placement report. The itemised list of where your money went is right there, and most accounts have never looked.
Leaving mobile app inventory switched on when you sell to businesses. Accidental clicks from games are the most common single leak.
Judging placements on click-through rate. A high rate on a game app means fingers, not interest.
Only reviewing the top twenty placements. The long tail is usually the bigger total.
Pasting in a hundred-thousand-line exclusion list you cannot audit, then wondering why reach collapsed.
Building exclusions at campaign level, so the next new campaign starts with none of them.
Migrating to Demand Gen before doing the audit, when the reporting was richer beforehand.
Using Display for cold prospecting on a small budget while Search still has room to grow.

FREQUENTLY ASKED QUESTIONS

How do I see where my Google Display ads are showing?

Open the placement report in Google Ads. It lists the websites, apps and channels your ads ran on, with the cost against each.

Sort by cost rather than impressions, and use a 90-day window. You’re asking where the money went, not where the impressions went.

Are Google Display campaigns being retired?

Yes, in effect. Google announced in May 2026 that standalone Display campaigns are being absorbed into Demand Gen.

A migration tool became available in June 2026. From January 2027, you can’t create new standalone Display campaigns, and remaining ones migrate automatically during 2027.

Display inventory isn’t disappearing. The campaign type that steers it is.

How do I exclude mobile apps from Google Ads?

Excluding adsenseformobileapps.com removes app inventory broadly, which is the quickest win for most business advertisers.

You can also exclude whole app categories such as games. Individual apps use the mobileapp::2- prefix followed by the app’s bundle ID.

Do it at the account level so it applies everywhere, and allow around 12 hours for it to take effect.

What is a made-for-advertising site?

A site built to earn advertising revenue rather than to be read. Thin content, very high ad density, articles split across several pages to multiply the ad slots.

The ANA found 21 per cent of programmatic impressions landed on this kind of inventory. Your ad appears, a human technically sees it, and almost nothing follows.

Are Display ads worth it for a small business?

For retargeting people who already know you, yes, and it’s one of the cheapest things you can do.

For finding brand new customers on a small budget, usually not. That’s where the waste concentrates, and it’s the use case most guides quietly assume.

How many placements can I exclude?

Up to 65,000 across as many as 20 lists at account level. That’s far beyond what a small account needs.

The real limit is how much of your placement report you’re willing to read, not the ceiling Google sets.

Will excluding placements hurt my campaign?

Your reach will fall, and your click volume will fall, usually by more than your spend does.

That’s the accidental clicks going away. Enquiries should hold steady or improve on less money.

If enquiries drop in proportion to spend, you cut something that was working. Look again at what you excluded, starting with any large downloaded list.

How often should I check the placement report?

Once a quarter is enough for most small accounts, once the obvious offenders are blocked at account level.

New inventory appears constantly, so this isn’t a job you finish. It just gets much faster after the first pass.

Ten minutes every three months keeps it under control.

Should I just download one of the big exclusion lists?

They’re useful as a second layer, not a first one.

They’re assembled from other people’s accounts, so some of what they block will have been fine for you. And a list that long is impossible to audit when something later stops working.

Start with your own placement report. That’s evidence from your own budget.

WHERE TO START THIS WEEK

Open the placement report. Ninety days, sorted by cost, highest first.

Read the top fifty rows and mark each one honestly. Could my customer plausibly have been here? Total the ones where the answer is no.

That number is the one worth knowing, and it’s yours rather than an industry average from an article.

Then do the two exclusions that take five minutes. App inventory if you sell to businesses, and anything from your own report that clearly doesn’t belong. Build them at account level so they apply to whatever comes next.

Do it before you migrate to Demand Gen, not after. The reporting is richer now than it will be later, and the exclusions you set at account level are the ones most likely to carry over.

Display isn’t broken. It’s just sold to you with the rubbish included, and nobody empties it unless you ask.

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: