Not Even Puppies Are Safe From Advertiser Blocklists

Illustration for article titled Not Even Puppies Are Safe From Advertiser Blocklists
Photo: Al Bello (Getty Images)

Even if you aren’t knee-deep in tech news all the time, chances are you’ve heard a tech critic mention the so-called “attention economy,” or as I like to call it, the “cash-for-eyeballs” economy. It’s a concept that, at least in theory, makes a lot of sense: Companies like Facebook make their multi-billion dollar empires, in large part, by hooking consumers onto a platform, which ensures that those same groups can see as many ads stuffed in front of their line of sight as humanely possible.

Advertisement

But just like IRL economy, the ongoing pandemic broke apart this digital economy, too. This is something you can really see with news sites, in particular—while a lot of us were seeing record-breaking numbers of people putting their eyeballs in front of stories about the ongoing pandemic, national protests, or the most bonkers election cycle in recent memory, those eyes weren’t necessarily translating into cash. The reason? Sweeping advertiser blacklists used to block their ads on anything a big brand wants to avoid.

I think we can all agree reading the news in 2020 was a process that skewed emotionally draining and mind-numbing at best. And it turns out, brands didn’t want their names being anywhere near these stories. And when stories using words like “covid-19" or “coronavirus” are too icky for brands to handle, they simply cut off their ads—and their ad dollars—from those stories. Back in 2019, the Guardian reported that brand safety tools were hemorrhaging around $3 billion dollars per year from publishers across the globe.

This problem didn’t start with the pandemic, and it won’t end with it, either. We’ve seen countless LGBT-focused publishers get brand-safety’d out of business over the years, and we’ve also seen countless stories get this treatment for leaning a bit too hard into politics.

Because the placement of digital ads is often automated, companies like to use lists of keywords that tip them off about whether a story is kosher as far as the “ick” factor is concerned. Many of the keywords include the obvious culprits—“violence,” “sex,” “porn,” “Rudy Giuliani.” But after getting my hands on a few of these lists myself, I can confirm that some of the “ickiest” stories on the web aren’t only those dealing with a global pandemic, but also those discussing... pets?

Some of the non-puppy related keywords we scraped during our research.
Some of the non-puppy related keywords we scraped during our research.
Screenshot: Shoshana Wodinsky (Gizmodo)

These lists came courtesy of Dr. Krzysztof Franaszek, the founder of the ad analytics platform Adalytics and the same researcher that helped me get to the bottom of a certain mysterious onesie ad.

Advertisement

The nitty-gritty of how he managed to find them in the first place is complicated, to say the least. He wrote out a full version for the coding nerds among us, but in short: After poking around in the developer console that comes baked into the Chrome browser, he noticed that the homepage for a particular news site (Reuters, in this case) seemed to be pinging endpoints with the names of a handful of major companies: Bank of America, Exxon, IBM, and the like.

After sifting through a few news stories with this same console open, he noticed that, buried in the code, many of these major names were labeled as “avoiding” some of the articles—and under that label, there were words: one story about pro-democracy activists in Hong Kong, for example, had the word “lawyer” marked under a list that seemed to be created for the Volkswagen brand. Franaszek will be the first to say this is “entirely observational,” but there’s enough evidence here to suggest that the car company is doing its darnedest to avoid running ads on anything vaguely lawyer-y. You can imagine that avoiding these sorts of stories may have come in handy when Dieselgate was still making headlines back in 2015.

Advertisement

Digging a little deeper showed that the list was coming from a company called “Admantx” that, back in 2019, was bought out by another company called Integral Ad Science (IAS)—itself one of the big names in the so-called “brand safety” business. After this, it was just a matter of finding the few dozen news sites that happen to be using Admantx’s tech, and scraping the lists of brands (and keywords) that cropped up under any given story. We’ve reached out to the IAS team to see if they plan to clamp down on this endpoint.

All told, Franaszek ended up with a list of dozens of Fortune 500 companies and other big brands—names like Walmart, Mastercard, Boeing, and uh, Smuckers (you know, the jam company). For the most part, the keywords these brands seemed to be blocking made some degree of sense. It makes sense that a company like JP Morgan wouldn’t want their name associated with a story that mentions “porn” or “beheadings” or “Al Qaeda.” It makes sense that Toyota doesn’t want to be seen with one of its car ads zooming next to a story about “the KKK” or “torture” or the word “dick.”

Advertisement

We’ve reached out to each of the brands named above and will update this piece when they respond.

Others are a bit harder to rationalize. Walmart, for example, didn’t seem to be avoiding stories using certain keywords as much as stories tackling particular categories: “religion and spirituality,” “food and drink,” and “pets” among them, meaning that your favorite puppy blog could (hypothetically) be on a blacklist. Some other brands also seemed dead-set on avoiding stories with their own sets of words that should be pretty innocuous: words like “kids,” or “truck,” among many, many others. Hell, there’s even a handful of familiar big-tech names like “Snowden,” Facebook chief product officer “Chris Cox,” and (naturally) “Elon Musk.”

Advertisement

All told, there were a little more than seven thousand keywords and phrases that we found here. By far, the most frequently mentioned were related to some aspect of violence—hell, “violence” was actually the most frequently banned word, followed closely by terms like “terrorism,” “crime,” and “murder.”

In some cases, brands avoided spending ad dollars on stories that used words like “investigation” or “California” or even the names of particular tech companies, like Amazon and Facebook. Again, I sort of get it—Facebook bad, after all—but that makes writing, say, an investigation on Facebook (or Amazon or Google or whatever) that much harder to monetize, which means news organizations might be a bit less inclined to pursue them at all.

Advertisement

It makes you wonder what kind of news cycle these brands are trying to leave us with.

Clarification: A previous version of this article stated that “digital ads are 95% automation.” While the process of placing ads is often largely automated, the 95% figure was meant to colloquially describe the process, not explain an exact figure. We’ve updated the story above to more generally describe the process.

Advertisement

I cover the business of data for Gizmodo. Send your worst tips to swodinsky@gizmodo.com.

DISCUSSION

I don’t know what the puppy in the picture did to anyone, but if anyone tries to hurt that precious lil’ nugget, I will personally punch them through the interwebz.