Feature

Cambridge Analytica scandal: what this means for adland

The incident has identified all kinds of tools and tactics that are relevant to the advertising industry.

Cambridge Analytica scandal: what this means for adland

The Cambridge Analytica scandal and its methods certainly piqued adland’s interest. If such powerful secret weapons existed, were they already something in marketers’ arsenals, and if not, could and should they be used for commercial purposes? 

Indeed, the IPA’s president, Sarah Golding, wrote in Campaign earlier this year: "The bad guys are simply using tools originally designed for the likes of Bed Bath & Beyond. I suggest it’s time we took them back."

An obvious point first – some of the tactics may not translate from political advertising to the commercial space. An election is an event and everyone is equal with one vote to "spend". 

Moreover, as Wylie points out: "The commercial space is not a zero-sum game. One per cent difference in an election means 100% of a win or 100% of a loss. In advertising, if you get 2% more customers in a year, that means 2% more revenue or market share. You don’t win the entire market with 2%."

Micro-targeting also does not lend itself to long-term brand consistency. That matters in the consumer space. While you might highlight different product attributes, it wouldn’t be wise to keep changing what a brand means to individual consumers by sending hyperpersonalised messages. A luxury car is a status symbol – and can command a price premium – only if everyone knows it is supposed to signal success.

Psychograhics – underused?

On the issue of psychographics, using the OCEAN model to categorise people by personality is "scientifically fairly robust", Rory Sutherland, Ogilvy’s vice-chairman, says. 

"Most people in psychology accept there are these five orthogonal variables. Most remain constant throughout life. They can even be observed in animals: you can test horses and hedgehogs for extraversion."

He believes most of adland would find themselves to be "disproportionately higher in openness" than the general population – something the industry should remember when trying to create campaigns that speak to a wide audience. But as for the need to use the technique, Sutherland argues that purely improving the creative on campaigns would often achieve better results.

Both Abbott Mead Vickers BBDO’s joint chief strategy officer Craig Mawdsley and Wunderman’s chief strategy officer, Richard Dunn, also see the use of full-blown psychographic profiling as unnecessary, and do not use it in their work.

"All segmentation is reductive and all marketing is targeting people’s psychology," Dunn says. "Nothing tells you the truth about what a person is, you are just trying to simplify how you cut your audience. It’s whether it is effective that counts."

With tools like "Lookalike Audiences", where Facebook will find people similar to those you want to reach, Mawdsley says the industry is already using behavioural-based targeting: "We’ve all moved on from crude demographics, but you don’t need to go to full psychographic profiling."

On the media side, however, Wavemaker UK’s chief product officer, Alex Steer, has a different view: "Cambridge Analytica’s approach is not unique... I know this, because my team do something similar… there’s very little voodoo about any of this."

The effectiveness of Cambridge Analytica’s techniques has been questioned. Did it really put Donald Trump in the White House as some have claimed, or did the company just have great PR? Or is the "liberal elite" just desperate to blame something for Brexit and Trump’s ascendance because they refuse to believe people hold those views independently? Both Cambridge Analytica and Trump’s campaign say that they did not use psychographic profiling on the election campaign.

Senior strategists are waiting to see the evidence before they make a decision. Sutherland believes the psychographic data based on the people who answered the questionnaire is probably quite accurate. But he is dubious about "the extent to which you can extrapolate people’s personalities from their Facebook ‘likes’".

Is it ethical?

Hanging over this debate is the question of ethics. Mawdsley likens the scandal to the controversy about subliminal advertising in the 1950s. It stems from the fear of being secretly manipulated. Adopting such techniques would be a dangerous game for adland, he says.

He adds: "People should always know when they are looking at an ad and when they are not. Advertising becoming a pervasive background force that works without your knowledge is not a good development for commercial life. 

"What we’ve witnessed is an erosion in trust in business and digital channels due to the feeling that people’s actions are being used against them. It’s potentially a very negative thing for the industry to get into. The clients we work with are very concerned about being trusted by their consumers."

Wise up on data

Wylie argues that adland needs to become more knowledgeable about data to avoid charlatans.

Sutherland agrees that there is a lack of statistical rigour in the industry. "There are far more data sets than people with the statistical analysis to [interpret] it. That’s dangerous," he says. 

"Statistics is an area where one rogue or spurious correlation can cause you to make decisions that are absolutely deranged on the assumption it is causation when it isn’t. The industry has a childish, slavish obsession with trying to make marketing a branch of data science. The lesson from my direct mail days was that the best creative approach worked better for everyone."

Read the full interview with Wylie