Friday, May 14, 2021

The AdNauseam test. | MIT Technology Review

Must read


Now positioned on both sides of an ad transaction, we were ready to observe the end-to-end lifecycle of an ad click. We invited individual volunteers to download AdNauseam and visit our site. We quickly recorded a few dozen successful clicks on AdNauseam, billed to our team’s advertiser account and credited to the publisher account. AdNauseam was working.

But this only proves that Google did not rule out the very first click on an ad generated by a brand new AdNauseam user recruited specifically for the test. To silence the skeptics, we needed to test whether Google would learn to recognize suspicious clicks over time.

We therefore conducted the experiment with people who had already used AdNauseam for some time. To anyone who watches for a very long time, these users stand out like a sore thumb, because with AdNauseam’s default settings, they seem to click 100% of the ads they see. Users can adjust the click-through rate, but even at 10% it would be well outside the norm; most people only click on display ads a fraction of 1% of the time. This test was therefore designed to verify whether Google would ignore AdNauseam clicks from a browser with a long history of astronomical click-through rates. If Google’s machine learning systems are so smart, they shouldn’t have any problem with this task.

An image of the AdNauseam “ad safe” collected by the Selenium automated browser.

ILLUSTRATION CREDIT: MUSHON ZER-AVIV

We tested this in two ways.

First, with people: we have recruited long-time AdNauseam users to access our website. We also invited new AdNauseam users to use the click software for a week during their normal web browsing, to establish a history, and then participate in the test.

Second, with software: We performed an automated test using a software tool called Selenium, which simulates human surfing behavior. Using Selenium, we ran a browser equipped with AdNauseam to automatically surf the web, browse sites and pages, pause, scroll, and click ads along the way. Basically, this allowed us to quickly create a record of prolific click activity while tightly controlling variables that might be relevant in determining whether or not Google is classifying a click as “genuine”. We configured four of these automated browsers and ran them for one, two, three, and seven days, respectively. At the end of each period, we sent browsers to our experimental site to see if AdSense accepted their clicks as legitimate. The Selenium browser that ran for seven days, for example, clicked on over 900 Google ads and almost 1,200 ads in all. If Google’s systems are indeed sensitive to suspicious click behavior, it should have set off alarm bells.

Most of our tests passed. Google filtered the clicks to our site through the automated browser which operated for three days. But it didn’t filter out the vast majority of other clicks, whether by regular AdNauseam users or even in higher-volume automated tests, where browsers clicked more than 100 Google ads per day. In short, Google’s advanced defenses were not responsive to the type of click behavior typical of AdNauseam use.

Google’s advanced defenses were not responsive to the type of click behavior typical of AdNauseam use.

Soon we had $ 100 in our AdSense account, enough to get Google to send us a check. We weren’t sure what to do with it. This money was by no means ill-gotten. We were just getting our own money back that we had put into the advertiser’s account, minus the 32% discount banked by Google. We decided not to cash the check. It was enough to know that we had proven that – for now, at least – AdNauseam works. The check was like a certificate of achievement.


However, our experience cannot answer other important questions. If you use AdNauseam, how do the clicks it generates affect the profile Google has built on you? Does AdNauseam succeed in preventing the individuals and populations in which they are classified from being targeted for advertising? (After all, even if you use the extension, Google can still collect massive amounts of data from your emails, search history, and other sources.) Even the answer to our simple initial question, to Finding out if the software works at all, took a lot of effort. To answer these other questions would require privileged access to many other online advertising nodes.

In fact, we can’t even conclusively know Why our test worked. Why Google did not detect these AdNauseam clicks. Was it a failure of skill or a failure of will?

Skill failure would mean that Google’s defenses against automated click-through advertising are less sophisticated than the company claims. However, as flattering as it may be to conclude that our small team has foiled one of the most powerful companies in history, it seems far-fetched.

A more likely explanation is a failure of the will. Google earns money for every click on an ad. If advertisers found out that they were being charged for fake clicks, it would of course undermine confidence in the online advertising industry. But advertisers can only validate these suspicions if they can look both sides of the market, as we have. And even if they could, Google’s market dominance it is difficult for them to take their things elsewhere.

In a statement, Google spokesperson Leslie Pitterson wrote: “We are detecting and filtering the vast majority of this bogus automated activity. Drawing conclusions from a small-scale experiment is not representative of Google’s advanced methods of detecting invalid traffic and the work in progress of our dedicated technology, policy and operations teams who battle it daily. advertising fraud. She added: “We are investing heavily in detecting incorrect traffic, including automated traffic from extensions like AdNauseum [sic]—To protect users, advertisers and publishers, as ad fraud affects everyone in the ecosystem, including Google. “

AdNauseam could adapt to circumvent Google’s counteroffensive, but an arms race will obviously favor Google.

If, contrary to Pitterson’s claims, the results of our experiment hold up on a large scale, it may be bad news for advertisers, but it is good news for internet users. This means that AdNauseam is one of the few tools that ordinary people currently have to guard against invasive profiling.

It is all the same a temporary and imperfect defense. If Google finds a way – or the will – to incapacitate AdNauseam, its usefulness might be short-lived. AdNauseam could adapt to circumvent Google’s counteroffensive, but an arms race will obviously favor Google.

- Advertisement -spot_img

More articles

LEAVE A REPLY

Please enter your comment!
Please enter your name here

- Advertisement -spot_img

Latest article