Data analysts need to take a look at themselves

As data analysts and scientists become among the most in-demand experts, Nick Booth wonders what would happen if they analysed their own use of data analytics

One of the joys of reporting on the telecoms industry was that it was full of uncomplicated, no nonsense characters who didn’t feel the need to invent their own special secret language – unlike the IT world.

Communications service providers (CSPs) were happy to connect people and let them get on with their lives. No dodgy software contracts packed with hidden licensing trojans. No built in obsolescence. No grandiose claims about saving the world, revolutionising business and bringing democracy to society. Granted, the CSPs and their dealers did charge huge amounts for managing moves and changes for their clients. But even that was a good discipline, as it stopped people moving about and over-complicating things.

The modern incarnation of the telecoms operator, the CSP, is mobile and data driven so, inevitably, they’re becoming more like IT companies. But, I’m relieved to say, they’re not quite the full Amazon, although many shareholders and equity holders won’t be pleased about that. They want their CSP investments to be omnipotent and ruthless.

Still, it’s obvious that data could save the life of the CSP by creating new opportunities, but then ruin it all over again by over-complicating things. The modern CSP has far too much complication to deal with, but they like to make life even more difficult for themselves by constantly moving their own goalposts.

The rapid growth by acquisition means that there seems to be a massive disconnect between the various departments of a modern CSP. The people in the call centre rarely have any connection with the intelligence available in their own data centres. John and Jane call centre operators only seem to be interested in two things. Getting you off the phone as soon as possible or selling you something you never knew you wanted.

There is a certain logic to this. According to David Zakkam, relationship head for data sciences company Mu Sigma, the average CSP could save a million dollars a year if it could shave a second off every support call. Zaikai advises 140 of the Fortune 500 companies on how to use data analytics to make each call more efficient and productive by giving call centre staff a better idea of who you are. However, having worked in a call centre myself, I suspect that many of the managers at shop floor level will take a much less sophisticated tactic. They’ll instruct their ‘here today, gone tomorrow’ minions to say anything to get the customer off the phone, once it’s obvious that they’re not going to buy anything.

Which brings us to the second option, up selling. Again, I will rely on the testimony on an expert witness from the telecoms industry, Eliano Marques, head of data science at Think Big. CSPs constantly offer things to clients that they have no use for at all. “I often try to understand the logic of the sales people at service providers. Why do they offer me 200 channels when it must be clear from their data that I only use ten. And why would they keep shoving a Samsung offer on me, when it must be obvious from their data that I’m an Apple fan,” says Marques.

It would seem there is a massive schism between the people who think for the company, the data scientists, and the people who act for it, in the call centres and the support teams. They seem to have completely different and counter productive targets.

Every CSP aims to progress with use cases that have an instant pay off, says Marques. Data scientists, whether internal or external, are hired to identify the best action for increasing profitability for the company, either through cutting costs or boosting the revenue. But the gap between potential value and the achievable value is often massive. For example, the best way to cut costs would be to create a predictive maintenance model that would identify where problems will occur and nip them in the bud. But it’s also the hardest. Partly because the people who are sent to do the fixing jobs have their own agendas, which don’t align with the company’s. The fastest way to meet your targets, I discovered, was to pass the buck, tick your box and re-assign. The thoughtful few who actually tackled the hard jobs would get no reward.

If only there was some way those people-centric problems could be analysed!