Mark Ritson: There’s a new weapon in the targeting armoury

Psychographic segmentation has been out of favour for a while despite its obvious usefulness, but combining it with behavioural and addressable data creates a potent mix that could make targeting supremely effective.

In many ways, 2016 seems like a lifetime ago. Cast your mind back to February 2016 and you will recall a very different place. We had the effortlessly cool and immensely reassuring figure of Barack Obama making most of us feel like the free world was in safe hands. The UK was an established and essential member of Europe and, aside from a few right-wing nutters, the future of the British Isles was inextricably and irrevocably entwined with our Euro-cousins.

And then it all changed. First, Brexit took this country by storm on that overcast day in June. A few months later, despite pussy-grabbing and wall-building his way through his election campaign, the Donald duly became the 45th president of the United States.

The common thread, other than incredulity, that links these two seismic events is a little-known British company called Cambridge Analytica. You may not know the name but you are about to become very familiar with the company in the months ahead. Led by impressive CEO Alexander Nix, the firm combines three disparate capabilities – psychometrics, big data and addressable marketing communications – into one apparently indomitable model used by both the Brexit and Trump campaigns.

READ MORE: Mark Ritson – Those who bash segmentation and targeting are talking nonsense

As any trained marketer can tell you, psychometrics is nothing new. There have been four options for segmentation for the past 50 years. You can segment a consumer market by geography, demographics, psychographics or behavioural means. Each has its own strengths and weaknesses and we often combine multiple modes together to create the most parsimonious approach.

Psychometrics makes a comeback

Until the recent arrival of Cambridge Analytica the use of psychographics in market segmentation had been in decline. That’s because psychographics has one big advantage and one even bigger drawback.

Clearly, if we know the outlook and psychological drivers of each consumer, we can create a powerful positioning message that speaks directly to how our product or service caters to their very specific outlook on life. If I am marketing car insurance to a ‘neurotic conscientious’ consumer, I push the idea of how well our brand will cover and compensate for any accident. If I am marketing the same brand to a ‘closed agreeable’ consumer, I emphasise the history and heritage of the brand and how it has become a tradition among better drivers.

We may be about to become as smart and as powerful as everybody in the market assumed we always were.

Despite the obvious advantage of being able to separate a market into groups based on their outlook on life and position them accordingly, the actionability of such an approach has always let it down. I might know, for example, from a representative survey that 28% of my target market are neurotic conscientious consumers. But unless I also know specifically the names and contact details of those three million people in the UK that make up this group, there is very little I can achieve with this segment because targeting is either impossible or highly inefficient.

That’s where big data comes in to play. Cambridge Analytica uses a well-established five-factor model called Ocean to psychographically profile a large sample of consumers into 32 predetermined sub-groups based on their answers to a series of questions about their preferences and motivations. Nothing new there. Where the firm went one step further was to also analyse these consumers’ other behaviours and attitudes gleaned from accessible secondary data such as Facebook activity, purchase history and so on.

READ MORE: Clive Humby – Social media can predict purchases as well as preferences

The power of secondary data here is that it enables the firm to extrapolate the psychographic segments from a small sample of a few thousand consumers to the entire population. As Nix told an audience at last year’s Concordia Summit: “We were able to form a model that could predict the personality of every single adult in the United States of America.”

Finally, with each and every adult identified and the respective campaigns developed to speak to the different types of consumer and their inherent personal motivations, the evolution of addressable advertising adds the third and final piece. With addressable advertising, Cambridge Analytica can send very different messages to individual consumers that they have already identified in their psychometric segmentation. This can be done through social media and also – as addressable TV advertising becomes a reality in the not so different future – via individual TV ads placed into programme content in real time via a set-top box.

Targeting’s newest tool

There is no doubt that this three-stage method was used by both Donald Trump in the final three months of his election campaign and by the Leave campaign in the UK to fashion the correct messages and deliver them to the right voters in the run up to the Brexit vote. Irrespective of involvement, the degree to which Cambridge Analytica was directly responsible for these two political shockwaves is a less certain point. Clearly, every success has many parents and it could be that this British firm is merely making the most of fortuitously finding itself on the right side of the polls twice in a row.

But a more cogent analysis of the Cambridge Analytica model suggests that it really was a major contributor to both political campaigns, perhaps decisively so. Its approach really does suggest a new and potentially more effective approach to communications than anything we have seen in the past. The confluence of big data, psychometrics and addressable advertising offers a potent new step in the evolution of marketing communications.

For marketers the implications are significant. Targeting has already become a complex and evolving discussion in marketing in recent years with a strong movement, led by Professor Byron Sharp, eschewing most targeted approaches for “sophisticated mass marketing”. This new approach presents a powerful alternative approach to that narrative and suggests that different messages, delivered to different targets at different times really might be back on the strategic agenda.

There are implications for the type of segmentation too. We have relegated psychographic segmentation in recent years and treated it as a poor cousin of behavioural segmentation that is built from market research. This new movement, and its ability to target everyone psychographically, opens it up as a fruitful and extremely interesting approach for many big brands to consider. The use of Ocean segments, all 32 of them, could well provide a more meaningful approach to segmentation than the banality of demographics, the complexity of behavioural segments or the incestuous limitations of the CRM system. We could see a renaissance in psychographics in the near future.

Marketers will face an inevitable backlash when consumers and activists discover exactly what is possible in terms of targeting and positioning. For many decades, consumers have assumed marketers were data-fed, manipulative experts who could change attitudes at the flick of a switch. The frankly piss-poor reality versus consumer paranoia always made me chuckle. If only, I used to say to myself, the public could see how fucking hopeless we actually are.

But with Cambridge Analytica and the combination of big data, addressable advertising and psychometric segments, we may be about to become as smart and as powerful as everybody in the market assumed we always were.

Professor Mark Ritson will be teaching the next class on the Marketing Week Mini MBA in Marketing from April 2017. To find out how it could make you a more confident, more effective and more inspired marketer, and to book your place, click here.

Hide Comments18 Show Comments
Comments
  • Mark Hodson 15 Feb 2017 at 8:47 am

    If they have really predicted the profiles of all ~300m Yanks, please get them to pre-identify the ~30,000 ones who will kill people: Minority Report style.
    Or isn’t it that accurate?

  • Jonathan Cahill 15 Feb 2017 at 8:53 am

    As Alexander Nix of Cambridge Analytica said “it will be interesting to see how they (these technologies) impact the next seven weeks”.

    It was. Cruz got just 19% of the votes at the Republican convention, Trump took 70%. As ever, it’s the bottom line which counts.

  • Tony Bond 15 Feb 2017 at 9:14 am

    “We were able to form a model that could predict the personality of every single adult in the United States of America” – really? Including people not using FB and paying cash for everything. I guess they may say they are all the same. But they’re not, are they? The most interesting element for an observer is the one you close with. Once consumers realise what analytics are being used here, will they change their behaviours? Personally, because I’m a bot arsey, I not infrequently fill things in wrongly, just because they’ve annoyed me. Maybe in the future, I won’t be so alone in this pastime……

  • sara 15 Feb 2017 at 10:04 am

    Very disappointed to see the lack of critical thinking on this piece that seems to have just gone with the PR release of Cambridge Analytica.

    For a good analysis I recommend this link: http://littleatoms.com/news-science/donald-trump-didnt-win-election-through-facebook

  • Zbiggy Ucinek 15 Feb 2017 at 4:00 pm

    Fully agree with the comments of Tony and Sara. Mark, without question, this was your worst article to date.

    Trump had an exceptional chance to win and Brexit was a real possibility. And no not being a post event smart arse, as my two American Democrat friends were told and vehemently argued against over 14 months ago, plus informing the then UK Foreign Secretary during a business forum re pre Brexit.

    It was obvious that anyone that has a knowledge of history, current affairs and a personal insight of real people, not so called useless models, psychobabble or bollocks, cover your arse research “because I have been taught to not have a clue to think independently or creatively” would understand.

    Issue – The political world gave us hope and provided none, we are struggling to survive.
    The Message – It all those bloody foreigners fault, not ours!
    Solution – Sod ’em and let’s try something new and screw all those politicians, this guy looks like he might mean business.
    The Result – Adolf Hitler

    Sorry Mark, stop defending or espousing a network and cartel that continues to sell the emperor’s new clothes. But then again it pays your bills and theirs so why should you be interested in changing it?

  • Piotr JJ Szymanski 16 Feb 2017 at 2:50 am

    According to this article https://motherboard.vice.com/en_us/article/how-our-likes-helped-trump-win , the way Alexander Nix and Cambridge Analytica came about their model (and, into fame) is dubious at most. Weapon or bullshit (Mark’s labels, not mine), the implications are scary — in marketing and in politics, to consumers and citizens.

  • Jonathan Cahill 16 Feb 2017 at 8:16 am

    On reflection, l think this is a good illustration of Scott Armstrong’s ‘seer-sucker theory’: “no matter how much evidence exists that seers do not exist, suckers will pay for the existence of seers.”

  • dinger 16 Feb 2017 at 8:21 pm

    cool, it’s neurotic conscientious paranoid reality Ritson bashing week – waddyamean, stormy overcast Brexit Day…

  • Lizette Nordin 17 Feb 2017 at 10:58 am

    Even if we believe that this level of sophisticated psychological profiling and behaviour prediction is possible, no-one ever seems to question whether we, as a comms and creative industry, have anything like the capability to develop the variety of personalised messaging required to make use of it. The crude car insurance example given above might work for a couple of segements but 32 or more different creative outputs? There simply isn’t the resource capability to produce this volume or variety in comms output to provide the kind of hyper-personalised persuasive creative that everyone gets excited about.

  • robert dry 17 Feb 2017 at 12:27 pm

    No doubting for me that the future will rely on the synthesis of different data points, the key is to start with the strategy, as mentioned in the article there is the Byron Sharp approach which still makes sense for high-scale international brands who can gain huge levels of scale from the broader approach. The highly targeted method will be useful for developing and growing brands, a practical example: I often see marketing presentations which are highly targeted in marketing comms but take a scatter gun approach to sales, going forward being able to identify your end user, where they will shop etc will help target distribution so you can focus efforts where there is a high likelihood of getting a sale.

  • Martin Hogan 19 Feb 2017 at 4:13 pm

    Strikes me this article is a bit like throwing a dart at the wall and then painting the bullseye around it.
    In my opinion, the next big thing in marketing will be a mixture of close analysis of behavioural segments allied to geolocation- in other words, getting a coherent message across at the time the target is closest to the point where they can actually respond.

  • Campbell Andrews 19 Feb 2017 at 10:43 pm

    Personalised data analysis and personalised targeting? Creepy plus. Anyone who values their privacy or dislikes the notion of a private Big Brother (Facebook, Apple, Google are all candidates) should be limiting their social media engagement and be very careful about what data is made available.

  • Ari Ginsberg 20 Feb 2017 at 1:22 am

    To take this further, Cognilyze have created the algorithms to turn purchase history alone into individual psychological profiles for each shopper.

  • Paul Rosham 20 Feb 2017 at 2:19 am

    I’d like to see evidence of the use of the platform and approach and understand what the campaigners actually did with the data.
    Two observations seem to complicate this as a something straightforward that I can accept as presented:
    1. Both the Brexit campaigners and Trump’s team seemed unprepared for what to do next. If they had an accurate predictive model and the foresight to know that it was going to be accurate, then their behaviours should have been more informed. Boris Johnson’s comments and the fact that Trump’s team are still disorganised indicate otherwise.
    2. The way that news item uptake and action seems to be more a factor of putting the messages into the correct channels over and over again. An illustration of the guidance that this data provided would be helpful. The counter proposal is that the Dunning-Kruger effect was the one that put them both over the top.

  • Rick Pullan 20 Feb 2017 at 6:32 pm

    will be interesting to see whether the whole model is thwarted by the ‘profiling’ restrictions of the new GDPR & ePrivacy laws.

  • John Douglas 20 Feb 2017 at 11:19 pm

    It sounds like a weapon. Using mega-data to observe measurable trends is simply scientific method, isn’t it. This simply provides the other side to the Marketing Science coin. It seems to me like the breadth vs depth debate just took a sizeable step beyond “I reckon” marketing.

  • Joshua Rex 12 Apr 2017 at 4:07 am

    We now know what we don’t know- and that is that, of CA’s very own admission, psychographic profiling had no hand and played no role in the work they did for Trump and that the Cruz campaign found their work useless and dumped them.

  • Post a comment

Latest from Marketing Week

PLEASE SIGN IN OR REGISTER. IT'S FREE, QUICK AND EASY!

Access Marketing Week’s wealth of insight, analysis and inspiration that will help you develop as a marketer and leader.

Register and receive the best content from the only title 100% dedicated to serving marketers' needs.

We’ll ask you just a few questions about what you do and where you work, so we can make Marketing Week more relevant to you.

Register now

THE BEST CONTENT

Our award winning editorial team and columnists will ask the biggest questions about the biggest issues on everything from strategy through to execution to help you navigate the fast moving modern marketing landscape.

THE BIGGEST ISSUES

From the opportunities and challenges of emerging technology to the need for greater effectiveness, from the challenge of measurement to building a marketing team fit for the future, we will be your guide.

PERSONAL AND PROFESSIONAL DEVELOPMENT

Information, inspiration and advice from the marketing world and beyond that will help you develop as a marketer and as a leader.

Dedicated to developing your skills and helping you achieve marketing excellence. Find guidance on leadership, professional development and the latest industry jobs.

Having problems?

Contact us on +44 (0)20 7292 3711 or email subscriptions@marketingweek.com

If you are looking for our Jobs site, please click here