Channel 4's latest 'Mirror on the Industry' report is diversity-by-spreadsheet from a clueless, self-appointed authority. Advertisers should reject it out of hand.
I think your points are well made and I don't disagree with your overall thesis. But (as a tedious planner type) I do need to point out a mathematical truth here. If 4% of the population is black, and the average ad has 5 people in it, then the correct, representative, percentage of ads that should have a black person in them is not 4%. It's [1-(0.96)^5] which is 18%. So black people are still over-represented but not by as much as you might think. On the other hand, South Asians are very much under-represented as around 32% of ads ought to be featuring them.
And of course, as you correctly point out, montage ads are seemingly unstoppable and frequently feature 10+ people. And following the same maths as above, 34% of ads with 10 people in them should feature a black person and almost 60% a south asian.
Which goes to show that advertisers are, of course, performing diversity by putting black people in ads because, to people who are detached from actual reality, diversity equals black people. As opposed to truthfully writing about Britain as it is, which would involve lots more asians in ads.
I should also point out that the writers of the report, who one might think ought to understand basic statistics, say that South Asians are well represented in ads because more than 8% of ads feature a South Asian. <eyeroll emoji>
Thanks Jacob, I think there's more sense in your comment than in the research – you should get a job at Tapestry 😅
The only things I'd add are that "average ad has 5 people" might be an unreliable baseline – I'm not sure how true it is and wish Tapestry would share that raw info. (Running this past ChatGPT, it's telling me something about Jensen's inequality and n = 5 being a generous assumption compared to n = ≥1. But I am now out of my comfort zone.)
Maybe more importantly, when an ad features five people, they're obviously not drawn as a random, neutral population sample. Most likely, they're meant to be a family round a breakfast table, a group playing football etc, which narrows the scope of potential casting decisions (unless you want a 75-year-old captaining the school football team etc) I think it's this obsession with assessing on an ad-by-ad basis that creates the weird incentives – if you're going to do it, then it's surely fairer to look at an advertiser's entire output over the course of a few years.
This comment puts me in mind of something I saw when working with Coca-Cola global 15 years ago. They'd shot a poster campaign showing families doing wholesome things and drinking coke. But they shot two versions of it: one with a black family and one with a white family. And they ran the posters according to the ethnic makeup of a neighbourhood. Which strikes me as a much more honest and coherent way to approach diversity, instead of cramming in one representative of each group into every ad. Of course most brands will never do this because of the cost issues, but still...
When I first came to this country (late 1990s) from Australia, I was struck at how clever & funny the ads were, often very short, plus that marvellous Guinness ad when we were at the pub watching the football. Feels like a different world.
You can see how the wild over representation happens though. It is pointed out to advertisers that a group is underrepresented. But advertisers aren’t pooling their data on how diverse their ads are; they’re independent and so they overcorrect like mad, and so you get a runaway effect. But once that’s in place, it’s very hard to be the advertiser who says “ummm, why have we cast this as mixed race?” because you might (almost surely will) get dinged by people who CARE about this. (Such as Channel 4, but also Bluesky.). So there’s a sort of deadly embrace or Reverse Prisoner’s Dilemma where nobody can defect because they’ll lose.
Yes well said - Rory Sutherland has talked about a kind of collective action problem, where every advertiser is making a rational decision but it leads to an irrational outcome.
I think what it doesn’t quite capture is the sheer level of monothink in the ad industry – none of the half-dozen trade press titles have covered any of this, and all push the same ‘we must do more’ narrative continually, all year round. So the initial decisions are only ‘rational’ in the sense that they’re responding to a false idea of public pressure.
The 'I am this guy' link is broken.
Story of my life. (Thanks, will fix it now)
Great article, by the way.
Very incisive and very funny - and a great title!
Thanks Sue, worked hard on that title 😅
I liked and unliked this just so I could like it again.
I think your points are well made and I don't disagree with your overall thesis. But (as a tedious planner type) I do need to point out a mathematical truth here. If 4% of the population is black, and the average ad has 5 people in it, then the correct, representative, percentage of ads that should have a black person in them is not 4%. It's [1-(0.96)^5] which is 18%. So black people are still over-represented but not by as much as you might think. On the other hand, South Asians are very much under-represented as around 32% of ads ought to be featuring them.
And of course, as you correctly point out, montage ads are seemingly unstoppable and frequently feature 10+ people. And following the same maths as above, 34% of ads with 10 people in them should feature a black person and almost 60% a south asian.
Which goes to show that advertisers are, of course, performing diversity by putting black people in ads because, to people who are detached from actual reality, diversity equals black people. As opposed to truthfully writing about Britain as it is, which would involve lots more asians in ads.
I should also point out that the writers of the report, who one might think ought to understand basic statistics, say that South Asians are well represented in ads because more than 8% of ads feature a South Asian. <eyeroll emoji>
Thanks Jacob, I think there's more sense in your comment than in the research – you should get a job at Tapestry 😅
The only things I'd add are that "average ad has 5 people" might be an unreliable baseline – I'm not sure how true it is and wish Tapestry would share that raw info. (Running this past ChatGPT, it's telling me something about Jensen's inequality and n = 5 being a generous assumption compared to n = ≥1. But I am now out of my comfort zone.)
Maybe more importantly, when an ad features five people, they're obviously not drawn as a random, neutral population sample. Most likely, they're meant to be a family round a breakfast table, a group playing football etc, which narrows the scope of potential casting decisions (unless you want a 75-year-old captaining the school football team etc) I think it's this obsession with assessing on an ad-by-ad basis that creates the weird incentives – if you're going to do it, then it's surely fairer to look at an advertiser's entire output over the course of a few years.
This comment puts me in mind of something I saw when working with Coca-Cola global 15 years ago. They'd shot a poster campaign showing families doing wholesome things and drinking coke. But they shot two versions of it: one with a black family and one with a white family. And they ran the posters according to the ethnic makeup of a neighbourhood. Which strikes me as a much more honest and coherent way to approach diversity, instead of cramming in one representative of each group into every ad. Of course most brands will never do this because of the cost issues, but still...
When I first came to this country (late 1990s) from Australia, I was struck at how clever & funny the ads were, often very short, plus that marvellous Guinness ad when we were at the pub watching the football. Feels like a different world.
Yes, sadly true
Great article. The stats will be for UK as a whole. Scotland's breakdown of ethnicity is a bit different. :)
You can see how the wild over representation happens though. It is pointed out to advertisers that a group is underrepresented. But advertisers aren’t pooling their data on how diverse their ads are; they’re independent and so they overcorrect like mad, and so you get a runaway effect. But once that’s in place, it’s very hard to be the advertiser who says “ummm, why have we cast this as mixed race?” because you might (almost surely will) get dinged by people who CARE about this. (Such as Channel 4, but also Bluesky.). So there’s a sort of deadly embrace or Reverse Prisoner’s Dilemma where nobody can defect because they’ll lose.
Yes well said - Rory Sutherland has talked about a kind of collective action problem, where every advertiser is making a rational decision but it leads to an irrational outcome.
I think what it doesn’t quite capture is the sheer level of monothink in the ad industry – none of the half-dozen trade press titles have covered any of this, and all push the same ‘we must do more’ narrative continually, all year round. So the initial decisions are only ‘rational’ in the sense that they’re responding to a false idea of public pressure.
Give someone a DEI hammer and suddenly everything starts to look like a gatekeeper’s nail. Excellent post Nick.