Earlier in my career, I was optimistic about the potential opportunities that opened up with objective programmatic advertising.
Biases based on business size, politics, race, gender, and culture could theoretically be weeded out by smart and objective robots or automation.
Experience has taught me that biases of the past won’t be going away anytime soon in digital advertising. They usually just take on new forms.
Technology has been used to increase efficiency and generate profits, but little has been done to use it as a weapon to weed out bias in digital advertising.
Billions of dollars in advertising flow through intermediaries such as Facebook, Google, and ad tech players such as Rubicon, Trade Desk, Index Exchange, and Open X.
Most of these platforms have some form of blacklisting and whitelisting.
Whitelists and blacklists have slightly different meanings depending on whether you are an advertiser or a publisher, according to Gamut. For advertisers, a blacklist is simply a list that identifies the sites on which an advertiser does not want its advertising to appear, while a whitelist identifies only the sites on which an advertiser wants its advertising to appear. You wouldn’t use both on the same campaign.
For publishers, similar reasoning applies but from the opposite side of the process. Publishers have whitelists designating which advertisers they will accept on their site and blacklists of advertisers that are blocked. The difference between the two is just the execution, block or allow. They achieve the same goal of regulating which advertisers are allowed onto the publisher’s site. The level of blocking can vary, from specific brands or landing page URLs to full advertiser categories.
Advertisers and publishers employ white and blacklists in a variety of ways and for their own distinct purposes. When used properly, white and blacklists can yield positive outcomes, but just as easily, improper use can produce negative results, according to Gamut.
Platforms such as Facebook and programmatic exchanges such as Index and Open X have whitelist and blacklist systems that drive how advertising is allocated within their platform. The platform itself can also whitelist certain publishers it wants to work with, and blacklist others.
There are many reasons why a site could be blacklisted on a platform. These include fraudulent bot traffic, too many ads on pages, or content that is racist, violent or sex oriented.
The platforms have a commercial interest in keeping their advertising neighborhoods safe and this is a smart business practice.
Having the policy of blacklisting some publishers and whitelisting others makes a lot of sense, but discriminatory issues could arise when when there is systematic bias and uneven policy enforcement.
Most of the decisions regarding blacklisting publishers are made by humans — white humans in most cases.
I have been in the digital media business for 11 years and I don’t remember ever meeting an African American who led or was involved in policy enforcement at any of the bigger platforms.
African Americans usually don’t have a seat at the table. An example of this is when Facebook cluelessly opened up the ability to block certain racial groups from ad campaigns, then removed it after complaints. This result was predictable with the right people in the room.
Facebook’s 2016 diversity report showed the monopolistic organization is only 2-percent Black.
The lack of diversity is important. Blacklisting some sites and whitelisting others has a material impact on how the annual $60 billion digital advertising dollars are distributed.
The Facebook platform applies blacklisting by seemingly arbitrarily deciding which publishers can use clickbait headlines (BuzzFeed) and which one’s can’t. Facebook has its own way of blacklisting and whitelisting media partners based on economic factors and other forms of bias.
The issue is not Facebook being more flexible with clients who spend a lot of money on its platform, but whether they are going to be consistent in their policy enforcement and whether their lack of diversity may create a commercial tax on media partners who fall outside the dominant group.
I have witnessed instances where Facebook just tells a publisher they are not allowed to market on its platform, but Facebook can’t explain why, and can’t explain why other sites are promiscuously allowed to abuse the same policies cited.
Can Facebook and other companies be trusted to be impartial in how they treat their media partners on the publisher and paid-marketing side?
When you allow a dominant group to drive and enforce policy, American history teaches us that the end result is massive and systematic bias and discrimination.
We saw this with Airbnb, where many Americans — left to their own choice — systematically discriminated against African American renters and owners.
We see the documented evidence over and over again in policing, housing, mortgage origination, consumer car loans, and business loans. We are already seeing bias in artificial intelligence behavior.
The issue with digital advertising is that disruptive trends are moving so fast there, that it is almost impossible for regulators to catch up to some discriminatory practices by some ad tech companies.
I have seen evidence using the Headerbid Expert tool — a plugin that enables publishers to maximize the monetization of their digital inventory — where celebrity news website TMZ (which has a lot of raunchy content) runs an exchange as a whitelisted publisher.
The no-bias argument is that TMZ has scale, and there is a commercial reason for uneven policy enforcement with the bigger ad tech platforms. However, I have seen multiple instances where African American-oriented publishers with good scale were “blacklisted” by the same exchanges endorsing a racy and raunchy TMZ.
There is another instance where Facebook said they didn’t want to allow a particular website marketing on their platform because some of their users complained. Did their users complain about several favorable articles about President Barack Obama? The evidence showed that users loved the site and content, but Facebook said the business was not allowed to do any marketing on their platform. This company pretty much owns a lot of internet real estate. Is there a possibility of digital redlining taking place?
Redlining is a discriminatory practice in real estate, typically involving lenders who refuse to lend money or extend credit to borrowers in certain areas of town — areas often occupied by people in poverty. It is against the law to discriminate against borrowers based on race or color, among other factors. That doesn’t stop people from trying.
Sure, my prior experiences and observations may not be statistically significant to prove a pattern of bias. However we do know that if there is any systematic bias in ad tech — as in other sectors of the economy — no one is watching and very few people, if any, care.
The companies will only change their behavior when regulators and lawsuits catch up, or there are material PR risks associated with their practices of bias.
Do you trust Facebook to be equitable with media partners when its own workforce is 2 percent Black? CEO Mark Zuckerberg had to stop Facebook employees from crossing out “Black Lives Matter” on the company chalkboard.