Big Data – People Analytics – Cell Phones – And Some Questions

Today, The Atlantic brings (back) our attention to the use of Big Data in the context of people analytics (“They’re Watching You at Work“) . Yes, I am all for evidence-based Human Resource practices. Leaving evaluation and judgement to machines comes with upsides – such as the reduction of hiring and promotion biases and a higher overall quality in HR processes on a company level. There are downsides, too. Let me pick one: privacy.

In the wake of Edward Snowden’s revelations of ongoing NSA practices – let’s consider this case and let’s assume it to be a hypothetical:

You are in a team of senior decision-makers in a telecommunication firm. You run a cell phone network. Your desire is to increase quality of cell phone service and you task a team of quality assurance and telecommunication engineers to sift through data to seek possible patterns among users who experience an above average number of dropped calls.

Makes sense, doesn’t it?

Someone on the team remarks that not all users are equal: Some only recently signed up for a 24 month contract and – despite annoying dropped calls – are more likely to hang in there than perhaps another user two months away from being able to switch providers without penalty. Data shows this is indeed the case. “Let’s throw a bonus to the at-risk customer to increase loyalty”, someone suggests.

Makes sense, doesn’t it?

“Well”, someone else involved suggests, “maybe not all users are equally finicky. Why not look at demographic data – our database can easily correlate that – and run an imperfect but decently accurate name-ethnicity or name-age or name-gender profiler to see it some groups are more at-risk of dropping their contract.” Data shows there are indeed demographic patterns. Some groups are more accepting of lower call quality than others. “Then let’s customize our discount/bonus offers to these groups respective ‘pain thresholds’ – saving us money and still reach the same level of customer retention.”

Makes sense, doesn’t it?

Feels right?