Marketers now know you better than you know yourself

By James Lumley, Business Writer for the South China Morning Post, the Irish Independent, the Sydney Morning Herald

Big data is an immensely powerful tool, but it has one big shortcoming: predictions based on past behaviour don’t help providers understand their customers.
 

We should embrace this innovation, but from the very start, we must ensure we do it responsibly.

When Jeff Bezos started an online bookstore called Amazon in his garage in 1994, he didn’t actually invent digital marketing, but he created what would become one of its most important building blocks.

Now, almost 25 years later, the company has a previously unimaginable database of client sales behaviour. Their algorithms can look at an individual’s previous purchases, map it against the herd, and predict with a high degree of accuracy what their customers want and when they want it.

As a result, Amazon is the world’s biggest online retailer by sales and market capitalisation, and Bezos is the world’s second-richest man. But this method, copied across the internet, has a serious limitation.

When Netflix recommends Downton Abbey over Breaking Bad, it doesn’t do so because it has any form of insight into that customer’s beliefs, outlook and personality. It does it because of statistics. Sometimes, it gets it horrendously wrong and consumers can be left with the feeling that Netflix just doesn’t “get” them any more.

The algorithm doesn’t even know why it is making a suggestion. So it can’t give marketers insight into their customers that can be used to develop produces or build campaigns.

Things are changing.

A recent study by Cambridge University looked into the psychological profiling possibilities of Facebook.

They proved that it is perfectly possible to generate a very accurate psychological profile of a Facebook user from their posts and likes using the so called “Big Five” personality traits. The same process can be used to analyse tweets, websites, and even emails.

The Big Five traits are: extraversion, agreeableness, openness, conscientiousness, and neuroticism. None of the traits are either good or bad.  They are simply descriptors. For years, psychologists have been testing people and rating each trait to work out individuals’ drivers and behavioural patterns.
 

What is new, however, is that it can now be done without making people sit a test. What’s more, the method is even more effective at working out a personality profile than the individual’s spouse.

But does it sell? The researchers decided to find out. They made two advertisements for a cosmetics product. One ad was targeted at consumers with extroverted traits, and the other was targeted at introverts. Online customers were shown the ad that matched their personality. Sales increased.
 

If one then takes the obvious next step and combines psychological profiling with big data, marketers not only know how likely a person is to buy the product, they also know how best to approach them.

But is this a good thing?

It certainly can be. As consumers, we are constantly being bombarded by companies that want our money. Would it be so bad if, in the future, we were only presented with things we like? Who could object to being shown a nice jumper or getting a good recommendation for a holiday read? That’s not intrusive. That’s customer service.

But what about companies that are selling financial products or investments?

Most of them will want to target people who are conservative-minded and responsible. Sensible, responsible people are more likely to take the time to consider their pension planning, or which investment they should make. This, again, is good customer service.

It could be taken a step further.
 

Psychometric testing, using carefully-designed quizzes, rather than social media posts, is already being used in developing economies to assess whether people without a credit history are creditworthy.

That isn’t necessary in developed economies, but there are many other ways in which the data could be harnessed.

For example, fund managers could be matched to clients based on what their psychological profile says about their attitude to risk.

A client with a low risk appetite could be looked after by a manager with a more conservative nature, while one more open to risk could be dealt with by somebody who is less risk-averse.

But follow the path too far, and danger lurks.

Some less reputable companies might want to sell their wares to people who are impulsive, and excessively risk-taking. That, too, could generate sales, but sales of low quality products to the detriment of the customer. 

It would cease to be customer service and become simple manipulation and mis-selling.

In the classic “boiler room” scam, fraudsters target the asset-rich elderly, and bully them into investments that turn out to be far too good to be true. Imagine how much more damage they could do if they had access to a psychological profile?

As consumers, unless we want to live in a hole under a rock with no mobile phone, we are all vulnerable. Social media, geo-locating mobile-phones, online shopping and old fashioned card spending creates all the data anybody could need to target us.

The most that we can do as consumers is to be aware it is happening, and recognise abuses. The recent “fake news” scandals shows that we should all be more questioning when we go online. But that doesn’t protect the most vulnerable.

For companies, it would be surprising for them to choose not to use information that helps them understand their customers better, and match people to products. But they, too, must be very much aware of the potential for manipulation.

With great power comes great responsibility. But in the end, it comes down to some of the oldest principles of business.
 

It doesn’t matter if one is selling cars, carrots, or complex derivatives, the same rules always have applied, and always will.