There is an interesting contrast between the methods used in the two ‘cousin’ fields of public relations (PR) and employee relations (ER). Whilst PR has become increasingly sophisticated in its use of technology, ER remains largely back in an era when digital technology meant using emails and offering ‘helpful facilities’ on the intranet
There has been no greater advance in PR’s use of digital technology than in politics. Obama’s second election heralded a radical turning point. It was then that the social media began to be utilized in a totally new way to get through to the electorate. With the aid of access to Facebook’s 2bn users Trump’s presidential campaign took this a stage further to exploit the repository of deep insights that can be used by those trying to influence social and political as well as economic behavior. When Facebook data is combined with purchasing choices gained by Amazon, searches on YouTube and social survey data it can not only allow Facebook to pinpoint messages that will be well received by its individual members, attract them to look at the messages through “click bait” – but then also precisely predict outcomes from PR/political campaigns.
The turning point for Facebook came when a US Professor of social psychology pointed out to them that the world was divided into what came to be called “Type 1” and “Type 2” people. Although its own internal technically sophisticated world (Type 2) was dominated by ‘logic’, ‘analysis’ and ‘probability’ around 85% of the population (Type 1) largely operate through gut feel, hunches, guesswork, prejudices and instinctive “just knowing”. The Type 1 population is also highly vulnerable to positive vs negative messages and emotional behaviours. People are generally stirred by powerful and energetic eloquence and take it as evidence of deeply held views, rather than often the insincere, contrived performance it often is. They are also taken in by false news – as long as it seems to be part of a contagion, replicated by several sources they know or respect – even if it has been orchestrated to appear so.
We may, as HR professionals, flinch at such a dichotomy – but it has proven to exist. That said, can we learn from such a finding when dealing with employees?
The usual value of social media for many HR professionals is to network with other business people via Linkedin, to search for potential candidates or to search for job applicant’s Facebook pages to see if they “check-out”. But what about using the social media to know more about employees and job applicants? During the Trump election Oxford Analytica was enlisted to work for the campaign as part of what was called Project Alimo. They found a way to use available data to predict many aspects of a voter’s personality, values and background. Systems would use data such as Facebook “Likes”, music preferences, expressions used, syntax and even spelling errors to predict Facebook personas – right down to a person’s religion and what subject they studied at University.
Data matching of this kind is almost certainly unlawful under data protection legislation in Europe, but the question remains – does data concerning, for instance, a person’s politics, religion or sexual orientation actually constitute “sensitive data” if it is not directly derived, and only predicted from other behaviours? Moreover, if the processing is conducted where it is lawful and users just ring in for information then no data protection infringement will have surely taken place?
Another huge advance made by analysts during the Trump campaign was in “sociometrics” – the evolution of groups and positions of individuals within them. Such an understanding is essential for the corporate world. To date it has only been possible to achieve reciprocated and unreciprocated bonding data by giving members of a group individual personality tests, or asking them to rate others in their group on a number of criteria. Many people object to such a rating exercise and personality tests can produce misleading results when it comes to converging their outcomes as they measure crude dispositions to bond in certain ways, not even the probability that bonds exist.
However, the kind of covert sociometrics that can be obtained from – say – Facebook browsing data and the interaction of Facebook friends can be remarkably accurate. For instance, the proportion of work colleagues that are in a person’s network and frequency of interaction with them can reveal a great deal – even if the actual contents of their communications is not known. Their likes and dislikes can also say a great deal about “value bonds”.
It is a matter of individual choice how far such insights are – or should be – sought and how they are used. Many people (like me) will reject them outright, others will fear their legality, but those who successfully exploit this brave new world of cyber-psychological possibilities will find themselves able to influence employee attitudes and behaviors in a way that is frighteningly effective.