NEWS13 March 2019

‘Unicorns’ needed to blend technical and human insight

Data analytics Impact 2019 News Technology UK

UK – Advancements in the research industry mean that new skillsets are needed that blend technical knowledge and human expertise, according to insights professionals.

The death of the question conf_crop

As companies increasingly employ data analytics alongside traditional research approaches, the insights industry will need to recruit talent who can bring both technical knowledge and human soft skills.

During a panel session at MRS Impact 2019, exploring how traditional methodologies are being merged with new analytical approaches, Leila Seith Hassan, data and analytics consultant, said: “The way I see it is it’s a shift away from a question, or the question, to asking lots and lots of different questions. Ten years ago, I would have to say: ‘I want a statistician’ and now it’s: ‘I want a data scientist’.”

It’s a challenge to recruit the right talent, added Hassan, because in addition to complex technical knowhow, they must have the “soft skills” needed to explain data analysis to those who don’t have that background. “We are asking for unicorns now, that’s what we need,” she said.

Increased automation won’t mean researchers losing their jobs, according to Hassan. “The technology, coding languages and algorithms are going to get more complicated and powerful but we’re always going to need people to intervene and steer those algorithms in the direction they need to go. We’re going to become more reliant on the people who speak the language of the technical but also speak the language of people.”

Human insight and research rigour is still needed to scrutinise data insight and seek deeper meaning in what is still a very experimental time for employing new techniques, according to the panel.

Claire Rainey, head of insight at O2, said: “If we’re not sure about what we’re getting back, we need to ask why.”

Alex North, head of marketing science solutions and partner development for EMEA at Facebook, said: “We as an industry aren’t really clear enough about how these techniques are being employed. The challenge is to provide the level of detail and scrutiny that can give confidence to the end user.”

Hassan said: “As we move to a world where everyone says ‘we’re using AI techniques’, I think everyone’s deploying almost black box techniques at the moment. These algorithms take in hundreds of variables on top of each other that come up with an answer, a number. We have techniques and we have answers, but we’re unable to say what happens in between.”

From Facebook’s perspective, the platform is placing more scrutiny over how it uses these approaches. North said: “It is important to be crystal clear about how you’re using people’s data. You need to be able to ask questions about these techniques. We should push back on the black box techniques. You are basically talking to a research team of a third party, so get them to convince you that their technique should be trusted.”