Save money AND flip off Google

There's a rising tide of anti-Google sentiment recently. And that's totally justified, considering advertisers have allowed Google to "grade their own homework" for years. Many advertisers have told me over the years that they desperately wanted to "get off of Google" but force of habit and other factors meant they just continued using what they were familiar with. After all "no one gets fired for using Google," right? Well, times are changing.

Recent high profile cases demonstrated again that Google is not even accurately tracking data on their own platforms and properly reporting discrepancies and problems to clients. For example, large amounts of video ads that did not conform to Google's own definition of "TrueView" were sold as TrueView ads [1 ]. YouTube ads that were not supposed to be shown to kids were shown on kids channels and kids videos [2 ]. Amidst these two recent scandals, it came to light that the "independent third party verification" vendors Google cited were not "independent" at all. They did not measure anything with javascript detection tags. Instead they were given data by Google and merely "performed calculations" and "provided reporting" which created the illusion of "independent measurement" but no ads were actually measured with their javascript tags.

It also became painfully apparent that these legacy fraud verification vendors often measure only 1 in 10 or 1 in 100 ads with a javascript tag (to save money, without telling their clients). That means 9 in 10 or 99 in 100 ads were not even measured, but still reported as highly viewable or low in fraud. That's just not right. I've already written another article -- Declared vs Detected -- Why Javascript Detection Tags are Crucial -- about why running javascript detection tags are crucial for accuracy of measurements, especially when it comes to fraud detection. Static .gif tracking pixels are unreliable for fraud detection, and also unreliable for basic counting, because any .gif pixel can be called by any browser or bot at any time. For example, video completion rates have been continuously gamed by bad guys because they can simply call the following tracking pixels -- firstQuartile.gif, midpoint.gif, thirdQuartile.gif, and complete.gif -- whenever they want, even if the video ad never ran. This is why you see video completion rates in the 90% range, even if your video ads were never served. Yeah, you should check on that.

Any reporting provided by the vendors you buy from is unreliable because they are grading their own homework and telling you everything is fine, so you keep spending money with them.

Without getting into any further details about why these vendors' tech sucks -- is unreliable or outright bad -- let me just focus on something very basic, that should be easy to understand. Note the screen shot below from 2017 when DoubleClick introduced a new way of counting impressions called "downloaded impressions." They wouldn't make such a change unless they had to -- i.e. they would be sued for fraud if they didn't. So what was the big deal? Read the yellow highlight below -- "count on download means that DCM (ad server) will count an impression when a user's device pings the creative download to BEGIN." Take a moment and think about this. "Pings the download to BEGIN" means the ad is being sent to the device; but it says nothing about whether the ad arrived in the device and rendered on screen. Take one more moment to think about this. DCM (now called CM360) is only reporting the count of ads they sent out, not the count of ads that actually arrived in the device and got displayed on screen. The latter is the more important count, if you were the advertiser, right?

Is this such a big deal -- the difference between ads served and ads displayed? Yep. If you ad didn't make it in time to be shown on screen, you won't get any adverting impact. The two slides below show the drop-offs between the second column (ads served) and third column (ads displayed). Even if the ad were served, it doesn't mean it got displayed on screen.

This is where a postbid javascript tag is crucial. For example, the FouAnalytics in-ad tags ride along with the ad. The FA tags are set to asynchronous, which means they fire when the ad finishes loading in the device. This way, the FouAnalytics in-ad tag serves as a good way to count the number of ads that got rendered on screen. By comparing this number to the number reported by the ad server, you can see whether you got what you paid for.

DCM/CM360 does NOT run a postbid javascript tag to determine of the ad made it to the device and rendered on screen. That is why their counts are unreliable. Furthermore, without a postbid javascript tag to collect data, DCM/CM360 doesn't even know where the ad actually ended up. It only takes the domain or app declared in the bid request and reports that, even if it were spoofed/faked by fraudsters. So the placement reports DCM/CM360 supplies to you are also unreliable in this respect.

So why did I waste my breath explaining all of the above?

FouAnalytics Barebones Ad Server of Record -- 1 cent display ads, 10 cents video ads

For all those who want "off Google" and an ad server that reports counts accurately (ads that rendered on screen, not ads sent out from the ad server) and placements accurately (where the ad actually went, as detected by a postbid javascript tag), we are onboarding a limited number of advertisers between now and the end of the year on my ad server -- FouAnalytics Barebones Ad Server of Record.

"I think there’s value in a cheap barebones ad server for clients who just need ad serving and reporting they can trust. And not be on google" -- ad industry practitioner

It is not fancy. It is meant to be barebones. But you can add ANY click tracker, attribution pixel, etc. you want so you can continue to use attribution platforms you are familiar with. This "ad server of record" does the basic job of 1) serving up display, video, HTML5 ads and 2) providing reliable counts and placement reports, simply because we run a postbid javascript tag to actually detect the data when Google's DCM/CM360 does not.

2023 Rate Card for initial advertisers on-boarded. These prices are grandfathered in for all time, from this point forward. Fees are paid up front, not in arrears. No payment terms like 30, 60, 90, 120, 365 days are offered. Advertisers buy a quantity of ads served up front and draw down from the quantity purchased.

• display ads - 1 cent CPM
• video ads - 10 cent CPM

DM me for more details. Save money AND flip off Google.