blog

Lies, damn lies and statistics

These famous words popularised by Mark Twain but commonly attributed to Benjamin Disraeli feel just as relevant today as then. Pertinent in looking at why, how and what we use data for as an industry – and as a global society.

Like me, Mark Twain, was actually a bit of a stats geek[i]. To give that some context I’ve recently been digging into the world of unconscious bias and data. Not sexy; but seriously important for understanding how we as marketeers may be creating or reinforcing bias through our clever use of digital targeting and data. It turns out that now, just as then, understanding the data you are collecting (or aren’t), it’s context and framing are everything. Our fascination with all things trackable is having an exponential impact on the variety and depth of data being collected[ii]. As deeper, more personal and institutional data becomes accessible, we need to look hard at the ethics of should we be using it. Individuals, both in work and life, share data without fully understanding how it can be used. Just measuring and comparing performance publicly can transform how sectors operate.

Take something as mainstream as University rankings; a way of globally comparing different institutions to assist the selection process for prospective students. Spearheaded by the Times Higher Education ranking, this has quickly come to define the sector. The reputational capital now generated has had an impact beyond the initial scope imagined. These rankings now actively influence the way Universities operate and their internal culture. For the simple reason that judgements can’t be made in ignorance of existing reputational hierarchies, all too often these types of comparison merely serve to reinforce earlier rankings. Creating more complexity simply amplifies this result. There are some strategies to address this, such as ranking subsets or regionalising data to show performance in key areas guiding prospective students in how they interpret the data. By deliberately framing this discovery in a meaningful way would help universities stand out vs their rivals. But understanding the impact of these types of data on your brands is still crucial.

With every consumer facing business now visibly reviewed and ranked on multiple social platforms (whether they want to be or not!); we increasingly need to consider dealing with these results as part of our comms planning. Ignoring it can be tempting; whether on tripadvisor, the app store or simply social commentary on your products & services, understanding the impact which these have on how people see and feel about your brand is important. Review the data and listen honestly, observing how it changes amongst different groups of customers. Identifying patterns whilst looking at what you can and can’t do about this is crucial. When discussing this with clients I always put myself in the consumer’s shoes. Reviewing how these reviews and social footprints interact with your owned and paid for communications is essential. Taking into consideration whether these either need to be addressed actively; monitored or, if positive whether they are something you deliberately seek to amplify and make more discoverable.

More ethically challenging, is the question of the data we as marketeers should be using to help influence what people are spending their time and money on. There is a fast- growing wealth of personal data available. By linking together people and data we can draw potentially useful conclusions about the ‘value’ of an individual. This can then be utilised for a particular purpose or highlight the level of risk they may represent. For clients and governments alike – even when trying to engineer better, healthier societies – the question becomes where to draw the line. With China moving now towards a social contribution score linked directly to which housing, education and benefits citizens can access at every point we have to ask ourselves that just because we can, does it mean we should?[iii]

Caroline Criado Perez explores a darker side of the biases in how these types of data are aggregated in her latest book[iv]. This sets out a compelling case for the role that the lack of gendered data for women plays in us viewing the world with a ‘default male view’. By not balancing and interrogating the data we collect, we not only reflect but sustain and even create continued gender biases in almost every aspect of our lives. Recalling Simone Beauvoirs’ The Second Sex (my introduction to feminism as a teenager), this book profoundly sets out the inherent cultural constructions which continue to reinforce this. Even in our very modern, digitally quantified world, the most striking examples were the role of un-gendered medical data, leading to a delay in the development of drugs and treatment profiles which work effectively for women. Everything from diagnosing heart attacks effectively in emergency rooms to the likelihood of men vs women being prescribed drugs vs counselling. This lack of good, balanced data for women holds back society as a whole. It is also profoundly relevant to our industry, for example Kantar’s recent research has highlighted that financial services providers have failed en-masse to cater to women. The fact that brands have failed to speak to women about finance in an appropriate way reflects on us as marketers and clients. By failing to collect and interpret data effectively and not understanding potential gender biases. In this case, a failure which has potentially cost financial services brands £130bn in unrealised opportunity- from 50% of the population. An audience who are increasingly seeking their own, independent financial provision with a brand who they believe will have their needs and best interests at heart[v].

In all of this; both commentaries really got me thinking about what this means for how much we use and rely on data to actually understand people. They act as a salient reminder to ensure that we understand the limitations and biases in the data we are using; and that we never stop asking the very human questions that allow us to create the emotional connections we know are crucial for effective brand building. We need to get into the real world and speak to actual human beings to get beyond the data to have any chance of generating insight that connects real people, through real emotion and real behaviour to your products. “What does the evidence say? Emotional appeals drive long-term effectiveness”[vi]. The sad flip side for marketing is that as this wealth of data has emerged, our marketing has become less effective. More short-term. Less impactful[vii]. That is what we need to overcome and ensure that we don’t simply mindlessly reinforce by only looking uncritically to our data for the answers. To value human insight as much as statistical analysis when we are trying to influence how people behave[viii]. To observe the biases and patterns that we may (unwittingly) be reinforcing with all our fancy algorithmic optimisations; and to understand whether that is something that supports or erodes your brand in the long term.

Want to know more?

We are talking about some of the issues in our Transformation Week, Brand Purpose session for clients on the 25th June.

As a passionate feminist, I am committed to changing this industry from the inside out. To see more women making their voices heard and leading in this industry. So in my spare time I’m working alongside a team of brilliant women to bring the highly acclaimed WACL Gather event to Edinburgh – the first Scottish event in its history. It’s been created specifically for female leaders & future leaders in advertising & marketing. Find out more & book your ticket at gatherscotland.com 

References:

[i] http://aaronjfisher.github.io/mark-twain-was-a-stats-fan.html

[ii] The Metric Society: On the Quantification of the Social (Stephen Mau, Polity Press, Cambridge, 2019).

[iii] https://www.forbes.com/sites/bernardmarr/2019/01/21/chinese-social-credit-score-utopian-big-data-bliss-or-black-mirror-on-steroids/#7811c76d48b8

[iv] Invisible Women: Exposing Data Bias in a World Designed for Men (Chatto & Windus, VINTAGE London, 2019)

[v] “UK financial institutions are failing to connect with female customers at every stage of the buying journey, from advertising to offerings. Advertising fails to communicate core tenets of ‘trustworthiness’, ‘understanding’, ‘dependability’, and ‘accessibility’ to women, and solicits positive responses from women less often than from men – according to analysis by Kantar’s facial recognition technology ‘Intuitive Association’1.

https://uk.kantar.com/consumer/shoppers/2017/financial-services-advertising-attract-female-clients/

You can download the full report here: http://www2.kantar.com/l/208642/2017-10-03/6cj9z

[vi] Anatomy of Effectiveness. A white paper by WARC, June 2019 https://www.warc.com/content/article/Anatomy_of_effectiveness_A_white_paper_by_WARC/126941

[vii] In a keynote address to delegates [at the Advertising Research Foundation’s (ARF) 2019 AUDIENCExSCIENCE conference], Binet asserted: “Of course, most marketers have learned completely the wrong lesson. They’ve seen the efficiency of short-term activation and they put all their money in there. But, in fact, what they actually should be doing is making digital activation work efficiently by supporting it with broad-reach, emotional brand building.” (For more, read WARC’s report: Les Binet examines how digital affects brand building/activation model.)

[viii] https://medium.com/@emmet.obriain/dehumanised-research-corporate-cultures-and-the-constraint-of-qualitative-insight-af00b0013ada

 

Next
Our internet picks of the week
Previous
Our internet picks of the week