Blue Christmas

Everyone v. Facebook, Jedi Blue, GrubHub, and Internet of SWAT

Everyone v Facebook

Facebook has been engaged in a…bizarre PR campaign against Apple for the last month. Apple announced it was going to require iPhone users to opt into app tracking that allows companies like Facebook to track users on other apps and the Internet. Facebook is claiming this would hurt small businesses, which is both not true and meant to conceal the fact it’d hurt them a lot:

Early next year, Apple plans to start requiring iPhone owners to explicitly choose whether to allow companies to track them across different apps, a practice that Facebook relies on to target ads and charge advertisers more.

Facebook has created a website and run full-page ads in newspapers talking about how mean and awful Apple is for giving their users a little more privacy.

How does this actually impact small businesses? That’s a good question. Facebook claims so many users will opt out of the app tracking - which would be a sign people don’t actually like it! - small businesses who are attempting to retarget ads to their customers won’t be able to do so. They say it will drive up ad costs. But will it? I don’t see any reason it would, and neither does anyone else:

Analysts said the change would probably have a limited impact on Facebook’s main ad business, as it already knows plenty about its users’ interests from their activity on Facebook and Instagram. But they said it could hurt Facebook’s efforts to sell ads in other places around the internet. Facebook’s enormous trove of user data is also one of its most valuable assets.

Before we get deep into display ads and retargeting, I want to talk about another recent revelation, one that actually does impact small businesses, in that they’re likely being ripped off by Facebook. A lawsuit revealed that Facebook’s employees are aware their interest and behavior targeting doesn’t work:

The documents feature internal Facebook communications in which managers appear to admit to major flaws in ad targeting capabilities, including that ads reached the intended audience less than half of the time and that data behind a targeting criterion was “all crap.”

Seems bad! Interest targeting, for those unfamiliar with Facebook’s ad platform, allows advertisers to target people based on things they are interested in, or behavior they’ve exhibited on the platform:

If you’re marketing a product to console gamers or ballet enthusiasts, these seem like pretty good options to reach your intended audience. But, it turns out, more than half the time you may not be reaching the people Facebook claims you are. They’re still happy to charge you as if you were, of course, and there’s no way for most advertisers to know.

The lawsuit cites other internal communication at Facebook, including this damning excerpt from 2016:

One engineer celebrated that detailed targeting accounted for “18% of total ads revenue,” and $14.8 million on June 17th alone. Using a smiley emoticon, an engineering manager responded, “Love this chart! Although if the most popular option is to combine interest and behavior, and we know for a fact our behavior is almost all crap, does this mean we are misleading advertiser [sic] a bit? :)” That manager proceeded to suggest further examination of top targeting criteria to “see if we are giving advertiser [sic] false hope.”

Facebook appears to know it’s on thin ice here, and has repeatedly sought to have the lawsuit dismissed, with some pretty revealing arguments:

The social network argued further, in its rejected motion to dismiss the suit, that it’s never guaranteed complete accuracy in its targeting, and that any claims of sophisticated targeting the plaintiff cited in its decision to buy Facebook ads were “generalized, promotional statements about Facebook’s advertising on which a reasonable consumer could not rely as guaranteeing a specific accuracy rate.”

As a grizzled industry veteran, I understand that there’s no way for Facebook to reasonably guarantee that all 117 million people in its ballet audience are season ticket holders, but the tooltips in its own ad platform say they are? It begs the question - is Facebook inflating audiences in its advertising system to seem impressive, or can it simply not target the correct customers, even with all its years of invasive data mining?

A core claim Facebook has made for years is that it has huge amounts of data about its users, and advertisers can take advantage of that to “microtarget” exactly the right people. It doesn’t sound quite as impressive to say that you have a 41% degree of confidence your behavioral targeting works.

The Pew Research Center conducted a survey in 2018 and found that 27% of respondents reported the data Facebook collected on them didn’t accurately represent them. More than half weren’t comfortable with Facebook compiling the information at all. Since 2018, Facebook has continued to erode the public’s trust in their platform, so it’s no wonder they are playing defense against Apple’s privacy changes and the bevy of lawsuits claiming their magical ad targeting is smoke and mirrors.

Jedi Blue

A couple weeks ago I wrote about the various lawsuits Google is facing. One of the arguments against Google is also an argument against Facebook, and deals with a contract between the companies, which prosecutors are arguing constitutes illegal price-fixing. When the lawsuits were filed they were heavily redacted but the Wall Street Journal got their hands on an unredacted version and it doesn’t look good:

In March 2017, Facebook publicly endorsed header bidding. Google approached Facebook and in September 2018 reached the digital advertising agreement, the states allege. The draft lawsuit says Google code-named it “Jedi Blue.” In December 2018, Facebook announced it was joining an advertising program, “Open Bidding,” that Google offers as an alternative to header bidding.

In return, the states allege in the final suit, Google gave Facebook special treatment. Among other things, it allowed Facebook to send bids directly into Google’s widely used software, known as an ad server, the draft lawsuit says.

What does all of this mean? Well, roughly speaking, Google set up a special arrangement with Facebook, giving it the opportunity to bid on ads it could match to its existing users, and gave Facebook favorable terms on the deal. In exchange, Facebook backed off endorsing “header bidding” which would allow website operators to offer their ad inventory to multiple exchanges - which Google was worried would jeopardize its monopoly.

So what’s the problem? Well, Google was giving Facebook first crack at targeting users with its ads, at less than half the price advertisers were paying on the Google ad exchange. And it didn’t charge Facebook for fake bot traffic:

Google also told Facebook which ad opportunities were likely produced by bots rather than by consumers, and it didn’t charge Facebook for those impressions, according to the draft suit. Other auction participants asked Google for the same information but were denied it, the final lawsuit says, redacting the information in question.

So, uh - how much fake traffic is Google knowingly selling to other advertisers without a special deal? Seems unethical at best! If you need a better explanation of how this whole ad ecosystem works, the graphic in the WSJ article is one of the better ones I’ve seen:

In conclusion - digital ads are bad, but they’re something the average consumer could probably live with if it meant a free Internet full of useful and entertaining things. What’s happened, however, is Facebook and Google have taken over, and are siphoning off huge fees for delivering the bad ads you can’t escape as you browse around. Fortunately - for the government - the companies were brazen in their pursuit of said monopolies, and left a paper trail.

Grubhub

In November, California voters passed Proposition 22, a gross piece of legislation crafted by gig companies and venture capitalists. Weeks later, Grubhub made a change to its app that drivers say has dramatically reduced the amount of tip income they receive:

Weeks after the ruling, Grubhub reduced its default tip amount from about 20 percent to zero, adding a suggestion to “leave an optional tip on top of driver benefits.”

Like other apps, Grubhub added an additional “benefit” fee, in its case $1.50, to each order in California—though that money is put into a centralized pot for which only a limited number of drivers are expected to fully qualify.

Not great! Drivers interviewed for the story reported a sharp dropoff in tips, which makes sense when gig companies are adding new fees to receipts. Consumers could be forgiven for thinking drivers are being paid better under the new law, but well, of course they aren’t:

Under Proposition 22, workers receive a healthcare stipend, provided they clock at least an average of 15 hours per week on one of the gig apps. However, in order to qualify, workers must already be the primary policyholder on an existing healthcare plan.

To get the full stipend, workers must put in at least 25 hours per week. The companies only count “engaged” time, not including periods spent driving without an assigned job — estimated to be about a third of all time spent on the road, according to a University of California, Berkeley, study. No allowances are made for time off or sickness. Data shared by Uber suggested that about three-quarters of its own drivers would not meet this threshold.

So generous! Gig companies will help pay for part of your health insurance, if you drive for more than 30 hours a week, assuming you can get enough gigs.

The true impact of Prop 22 on gig workers in California will take months and maybe years to be fully measured and understood, but Grubhub has gotten out in front of the pack by going after tips, estimated to make up around 30% of a driver’s earnings.

Internet of SWAT

The FBI has issued a warning to users with “smart” home devices with camera and voice capabilities. People tend to leave default passwords set on their devices or use common log-ins, and now hackers are taking advantage of this and “swatting” people - placing hoax calls to the cops and reporting threats that result in an armed response. Swatting is incredibly dangerous as US police are trigger happy and suffer no consequences for the murder of unarmed innocents. The best option is to not buy internet-connected home devices to begin with, but the second best option is to use a unique password or two-factor authentication. Technology kicks ass.

Short Cons

Reuters - “Switzerland will return $150 million from blocked Swiss bank accounts by the end of the year to the United States to be given to victims of convicted Ponzi scheme con artist Robert Allen Stanford…”

ProPublica - “Alice Stebbins was hired to fix the finances of California’s powerful utility regulator. She was fired after finding $200 million for the state’s deaf, blind and poor residents was missing.”

Business Insider - “President Donald Trump's most powerful advisor, Jared Kushner, approved the creation of a campaign shell company that secretly paid the president's family members and spent almost half of the campaign's $1.26 billion war chest...”

Tips, thoughts, or billion dollar war chests to scammerdarkly@gmail.com