Okay so that’s some serious hyperbole. But so far, this little trick is working for every account my agency has tested it on.

The Click Fraud Problem

Click fraud can be a nightmare for PPC advertisers. Not only does it eat up budget that should otherwise be used to drive leads and sales. It can also destroy the performance of campaigns when it includes fake lead generation.

google ads click fraud

Here’s a brief list of some common tactics the fraudsters use:

  • Botnets
  • Click farms
  • Invisible ad placements
  • App hijacking
  • Click wars
  • Click bombing
  • Brand name sabotage

How to Identify Click Fraud

Unfortunately, click fraud happens on the GDN (Google Display Network), YouTube Ads, App, Discovery, Search, & Search Partners networks. Yep, that’s all of them!

Often advertisers only discover click fraud because campaign performance tanks.

In the case of general click fraud you might notice that performance drops off on particular campaigns. On display, for example, you might identify that you’re suddenly getting a lot of click volume from a few websites or videos. When you check those out they are clearly garbage placements yet have extremely high click through rates. Over on search, you might see that you’re getting most of your clicks from the same few search terms. Again, you might notice an extremely high click through rate.

In some cases you might also suffer from fake leads. First, you may notice a sudden big increase in campaign “performance” with many more conversions than before. Second, you might find that while you seemingly are getting more conversions, most of those are fake leads or invalid conversions, e.g. many “click to call” style conversions with no actual calls.

In these cases, you will not only notice a very high CTR on the offending placements or search terms, but a similarly high conversion rate.

Traditional Methods to Stop Click Fraud

I wrote a long article about click fraud and click fraud protection services. I go into a lot of detail there about different types of click fraud and what you can do to try to stop it. Here’s a grocery list:

  • IP exclusions
  • content exclusions
  • search term exclusions
  • geo exclusions
  • geo or bot blocking with Cloudflare or similar
  • open an investigation with Google support
  • install 3rd party click fraud software
  • use CAPTCHAs
  • use Honey Pots
  • track offline conversions exclusively
  • stop using the offending channels

Unfortunately, while some of these steps can stop click fraud temporarily, sophisticated schemes can get around most if not all of these measures, outside of stopping your ads. Well and often “stopping your ads” is exactly what the fraudster (in the case of a competitor) was trying to get you to do in the first place.

Stumbling Upon an Interesting Data Anomaly

Back around the end of October 2025 I noticed a trend in one of our client’s search campaigns. The CPA had dropped like a rock over just a few weeks… whereas it came in between $40-70 for years, it was suddenly turning in $5 CPA figures.

I knew something was wrong and when I looked through the Unbounce contacts I could see something was up. The business names given by leads were all massive IT companies like Twitter, Google, Microsoft, and so on. This doesn’t fit the customer profile for this advertiser so my assumptions were confirmed.

Next I started to look at search term data and could see clearly that there was an abnormally high volume of low relevance search terms driving super high click through and conversion rates.

I immediately blocked those queries with negatives but of course within 24-hours the fraudsters got around that with new queries.

Digging further into reports I noticed something very interesting. When looking at demographics I could see that the vast majority of conversions were coming from the unknown segments. This was true across age, gender, and income level.

Excluding Unknown Demographics

With this new information in hand I excluded the “unknown” gender demographic across all campaign ad groups. And voila, just like magic, all click fraud ceased immediately and 10-months later hasn’t ever returned. While I could have also excluded unknown ages and income levels, that’s really unnecessary… fraudster/user clicks won’t include any demographics data.

Since this example, we’ve implemented this for a half-dozen other clients experiencing paid search click fraud. It has worked in each and every case. Oh and it also works over on Microsoft Ads too.

How & Why Does it Work?

Here’s exactly what I believe is going on under the hood.

For a particular user to be associated with demographics, they must be signed-into a Google service. Google combines direct data from signed-in Google Accounts with inferred data gathered from web browsing activity, app usage, and geographic location signals.

There are a few things to dissect here:

  1. For a user to include demographics data they must be signed into a Google service
  2. Users that don’t declare demographics directly in Google accounts will require activity (read time) before Google can infer demographics

Sophisticated fraudsters use bots with VPNs (rotating IPs) to generate search queries, click through search ads, and fill out web forms or other conversion actions on site. Because of this they won’t ever be signed into a Google service or have any preset or inferred demographics data.

Could they get around this issue?

Perhaps they could if they generated a bunch of fake Google accounts and populated them with demographics data. And then logged into those individual Google service accounts within the VPN/bot account.

But that’s complicated and of course Google doesn’t like it when you suddenly “show up” on another device or from a new location; requiring a fresh log in with 2FA. On top of that, just consider how many Google accounts fraudsters would need to create and manage. While accounts could be recycled for clicks, they couldn’t be for the same target advertiser. It’s also not clear to me whether Google will automatically establish demographics for an account that declares age, gender and address (for income), or whether Google still requires browsing activity before setting those. Or what the time frame is before an account gets assigned demographics.

Bottom line this is a huge amount of overhead and additional cost, if even possible, probably making “click fraud for sale” too expensive to justify when there’s lower hanging fruit.

Why Not Exclude Unknown Demographics by Default?

Despite how well this has worked, we do not exclude unknown demographics by default for our clients. For most advertisers, unknown demographics users make up around 30% (rough average estimate) of impressions, clicks, and conversions. That’s a lot of search inventory to cut on speculation that you might be attacked in the future. Sure, I suppose if you’re only bidding on a very small impression share of total inventory, this might be an okay move.

But generally I’d wait until you can evaluate performance before considering carving off so much potential traffic.

Does this Work for Other Channels too?

Yes, this seems to work across all Google Ads channels, noting we haven’t performed rigorous testing yet. If the campaign type supports excluding demographics you can prevent click fraud for that campaign using this trick. Importantly P-Max does now support hard demographics exclusions in campaign settings; not to be confused with “soft” exclusions in audience signals.

However, click fraud and low quality clicks/conversions are different things. This trick will not prevent:

  1. Real users from clicking on your ads and converting, even if they aren’t seriously interested in what you have for sale
  2. Kids from clicking on your video ads signaling to Google that it should run children’s videos exclusively for your P-Max or video campaign
  3. Accidental clicks on app ads, especially for gaming apps
  4. Bad conversion setups on your website such as click to call or click to email that send bad signals to the bidding algorithm

Summary

Click fraud is a very serious problem for many PPC advertisers. Until now there hasn’t been a simple universal fix other than stopping offending campaigns completely. This new trick seems to work effectively and keeps click fraud at bay for the long term. However, fraudsters are constantly evolving their toolbox and may overcome unknown demographics blocking at some point.

Still, combining this with other tactics in the future may make the cost of click fraud too expensive to make sense. Anything we can do to reduce this scourge is welcome!