Anonymous.google – A Problem Child? – Updated

Network: Google Display Network

August 2015

Name: Anonymous.Google

Class: GDN-Anonymous-Publishers

Dates of Traffic: March 2014 – May 2015

Traffic Volume:  1.8% of US-Only Display, 2% of International

Publisher Volume: 4% of US-Only Display, 2% of International

Publish Date: August 19th 2015

Announcement Date: August 19th 2015

 

Summary

Google allows publishers in the Display Network to remain anonymous in the reporting available to advertisers through AdWords. In its official documentation, there is very little explanation of the reasons for allowing anonymous placements. The lack of transparency, and the inability to gauge the appropriateness of the context of placements on anonymized sites, presents advertisers with both a serious brand risk — that of having ads shown on inappropriate sites — as well as raising questions about the quality of clicks generated from those placements.

Anonymous Google placements accounted for about 2% of all billed traffic in our surveys of the Display Network, whether targeting US-Only placements or International. In US-Only tests, anonymous placements generated nearly twice the percentage of robotic traffic as to non-anonymous placements, and while we believe that Google did not bill for the majority of those automated clicks, their presence raises suspicions about anonymous placements in general.

Why Do Anonymous Placements Exist?

We can no longer find any pages within Google’s support documentation that explain the rationale behind anonymous placements, but various articles contain excerpts from a previously existing support page, which explain that:

“You might see placements marked as “123456.anonymous.google” in your results. Some publishers choose to offer placements anonymously and not disclose their site names to advertisers. Even though the publisher and website remain anonymous, you will still have access to the reporting statistics from the ad placement.”

This doesn’t tell us much, but we may glean that it’s a service intended to be to the benefit of the publisher, rather than the advertiser, and that it’s not a question of every click from a publisher being anonymized, but some selection of possible placements. Why would a publisher which to remain anonymous? We speculate that it might involve arrangements where the publisher is filling ads on otherwise unsold impressions at a lower CPC, or for less desirable keywords, and in exchange requests anonymity. In any case, advertisers should be aware that anonymous targeting exists to benefit the publisher, and not the advertiser.

Identifying Anonymous Publishers

Our speculations are in part based on observations that allow us to identify the actual publisher sites behind a particular anonymous.google domain. In AdWords, advertisers can append special tracking parameters to the target URLs associated with an ad. These are called ValueTrack parameters, and they allow the advertiser to glean information about the ad placement and associate them with the specific visitor session associated with he ad click. Information on ValueTrack is available through the AdWords support site.

The ValueTrack parameter that is relevant to assessing anonymous.google traffic is referred to as the “placement” parameter. When a visitor is redirected to the advertiser site following an ad click, the “placement” parameter specified in the page request, filled in by AdWords, contains the domain of the publisher site on which the ad was served. For anonymous placements, the value in this parameter is the same anonymous.google domain under which the click will be reported in AdWords.

While Google can anonymize the “placement” parameter in the request, the user’s browser, unless configured by the user to suppress referring URL headers, provides the real URL where the ad was shown, allowing us to identify the real placement and draw some conclusions about anonymous placements:

  1. Anonymous site identifiers are constant. For example, e6455cf004ccda74.anonymous.google is always speedtest.net in our surveys, with clicks spanning more than a year over multiple campaigns. It is possible to evaluate the performance of anonymized domains over time, since they are reported consistently.
  2. Not every paid click from a given publisher is anonymized. There are anonymized publishers from whom we’ve been billed clicks that we’re reported under the real domain of the publisher. As of yet, we do not have enough data to draw a statistically meaningful conclusion about whether anonymized clicks from a given publisher occurred at a lower average CPC, although our current data supports the possibility.
  3. There are large anonymous publishers you may recognize: tomshardware.com, about.com, cnet.com, and reference.com. The number of larger anonymous sites, in comparison to the non-anonymized traffic, is what suggests to us that anonymous.google may be mostly about impression inventory management for higher traffic publishers.

Evaluating Anonymous.Google Traffic Quality

The first thing that stands out to us is the staggering amount of simple robotic clicks coming from anonymized placements in US-Only Display. The following charts show the results of PureCaptcha validation on all incoming traffic from the ads in our tests, regardless of whether it was billed:

Anonymous.google invalid traffic

 

While Google does not provide enough information to reconcile actual traffic with billing data on a per-click level, we are reasonably confident that the majority of clicks filtered as “identified bots” in these particular tests were not billed.

However, the sheer volume of simple bot traffic coming through these ads causes suspicion. In US-Only tests, as a whole anonymous placements drive a lower quality of traffic than non-anonymous placements.

In International traffic the overall volume is less, but the trend of increased simple bot activity on anonymous placements was consistent, with three times as much robotic traffic from anonymous.google placement as from other international placements:

 

anonymous.google

 

Evaluating Anonymous.Google User Behavior

Our second comparison filters out the identified bots and duplicate clicks in order to look at the distribution of user behavior for clicks that pass our first set of IAB filters. As we’ve seen in our publications about unresolved traffic, user behavior distribution in a click population can be a significant indicator of more sophisticated forms of fraud. With anonymous.google, once we eliminate the large numbers of simple automated clicks, the remaining traffic appears similar to the rest of the network:

Behaviors

international anonymous.google

Is Anonymous.Google Risky for Advertisers?

The data gives a complex answer to this question. The fact that so much of the raw traffic coming from anonymous placements is made up of simple bots is alarming. The fact that Google already identifies and filters the simplest of those bot clicks helps, but we wonder why they continue to tolerate publishers sending such a high volume of fraudulent traffic, even if it is filterable. There is certainly a risk that in tolerating simple fraud because it is identifiable, they leave the door open for more sophisticated and less easily caught forms of fraud.

On the other hand, once that traffic is removed from the data set, the remaining traffic looks very similar to the rest of the network. This is reasonable given what we know about anonymous placements. These are Display Network sites that also generate non-anonymous placements some of the time. Overall, the network quality problems with resolution and user confirmation of intent exist for both anonymous and non-anonymous placements. In terms of traffic quality, we judge the risk to advertisers of anonymous traffic to be slightly higher than that of the Display Network as a whole.

Beyond user behavior and click validation however, there is the question of brand risk. If you don’t know where your placement is being served, how do you know it’s a good context for your brand? Taken together, the possible brand risk and relatively poor traffic quality suggest that advertisers may wish to avoid anonymous Google traffic.

A Final Note on Transparency

The Internet offers much to advertisers. It offers a higher degree of transparency compared to traditional media and if it is to attract more advertising dollars, it should continue to embrace the ideals about transparency that got it to this point rather than hiding in the shadows.

Those who perpetrate click fraud benefit from a lack of transparency, whether it takes the form of the difficulty advertisers face in attempting to reconcile website traffic with billed clicks, or the question of anonymous publishers.

We believe that the existence of anonymous publishers reflects an unhealthy bias in the industry towards the interests of publishers over the interests of advertisers. Technologies which allow for click resolution and post-click analysis, as well as increased transparency about publishers would be beneficial in swinging the balance back in a more positive direction for advertisers and the industry.

How to block Anonymous.google placements

In Google AdWords

1.  Open your Google AdWords Account:

2.  Click on the Campaign Tab

3.  Select the Display Network

4.  Select Placements and scroll to the bottom of the screen

5.  Click on the exclusions word with the plus sign next to it.  Put “anonymous.google” in this box and follow the directions, be sure to click save.