Earlier this year Facebook stopped allowing American advertisers in key categories to show their messages only to people of a certain race, gender, or age group. One of these categories is jobs. This is of course in compliance with discrimination laws that I don’t think anybody will argue with. The problem here is not the law itself, but the solution that is used to obey by this law, that actually leads to exclusion by design.
The problem with this and solutions currently on the table, some of which are actually implemented in certain countries already, is that with all the right intentions the opposite is achieved.
Difference With Paper Advertising
I know many people who say the same things happens with paper advertising. If you place the vacancy in a young girls’ magazine, you will not reach (old) men. If you place the vacancy in a retirement magazine, you will not reach the young.
But social media is different, when you only allow ads to be shown to a specific group. They don’t see the ad at all, while they may see the magazine on the rack. Even though you are not targeting a specific group when advertising in a magazine, you’re not excluding them from being able to see it either. With social media targeting, you are.
The Effect of Targeting
The problem with the new system of not allowing specific targeting is that you now also need one ad for all groups. You cannot tailor the message or the images used and that will lead to exclusion by design.
Some great examples come from an award-winning campaign to recruit security staff at Amsterdam Airport Schiphol. Its success was build on the personalization of both the visual as well as the text and the landing pages of the ads served mainly on Facebook.
It found that people with what we call “a non-western migration background” were under the impression they would not be welcome, even though the opposite was the case. It also found that the main motivation for wanting a job at airport security within this group differed by age. The younger men loved the good pay. The older men loved it because they looked good in a uniform. So they build different ads to cater to both categories, targeting on age, with a different message and a different person on the photo.
The targeting by ethnic background was done using people with interests (likes) specific to those backgrounds. Of course the native Dutch were not excluded, but they were served a different ad, with a photo that looked like them, whether male of female, with again copy suited to their reasons for wanting this job. Women in their thirties and forties loved this job because of the flexible working hours and the fact that, unlike in Dutch healthcare, schedules are not made just a week in advance. So they were targeted with ads showing a women and copy that emphasized the flexibility and the part-time working opportunities.
The effect was amazing. It ended up hiring 1,513 people in 14 months at a cost lower than ever before at this organization.
On top of that, the organization has become much more diverse with regard to both ethnic backgrounds as well as age. Because of the campaign, there are now more than 80 different social-cultural groups working at the organization.
It did not exclude a single category of people, but did serve different messages to different groups, making every group feel welcome.
If you cannot target, you cannot do that.
One Ad to Rule Them All
The problem we are now facing with the inability to target job ads based on gender and age is that we need to create one ad for all. Let’s call it the vanilla ad; something that’s just average. And the average is usually a white male in his thirties. So people of color will feel excluded seeing the ad. The opposite is true if you’d put a photo of a minority man in the ad.
And let’s not forget that women look at ads with women more than those with men. That’s genetically wired in our brains. We trust people that look like us more so we are attracted to them.
By not allowing targeted advertising, we are building advertising that is exclusive, not inclusive, by design.
Another problem is budget allocation. Suppose I work at a major IT organization and I want to hire more female engineers. I am not allowed to just target females and since I don’t want to discriminate against men, that’s good. However, if I use one campaign, I have a second problem next to the “vanilla ad” problem.
I would like my budget to be spent equally on men and women. The problem is I cannot target and since 80 percent of all engineers are men, 80 percent of my budget will be spend on men. This way I will never change the status quo at my organization.
The solution would be to allow targeting on age, gender, and even race, but demand that all need to be included in a campaign. So the campaign needs to be inclusive, but within a campaign targeting specific ads to specific groups is allowed, even based on age, gender, and race.
Of course there are ways to use this for the wrong reasons. By lowering the budget for specific groups you could technically exclude them, or you could exclude them by making poor ads for these groups. If an organization wants to discriminate, it’s hard to fight that with technology. You might be able to use their technological discrimination to fight them in court.
But the way we are moving to right now is by building systems that seem inclusive, but are excluding people by design, literally.