On data scientists, assorted weirdos and empathy

  • Sam Menter

After the election it became clear that Dominic Cummings would be playing a significant role in government working for Boris Johnson.

Cummings was campaign director for Vote Leave and has been working in and around government for the last 20 years.

He has a reputation for being a maverick and a disruptor, he’s written about transforming Britain into a “meritocratic technopolis“; and is a longstanding critic of the civil service which he wants to restructure.

Cummings is recruiting help. His job descriptions paint Number 10 as an innovative start-up ready to ‘fail harder’ rather than as one of the UK’s most established institutions whose decisions affect us all.

Exciting but slightly terrifying. This is the age we live in.

I’m encouraged by the fact that he quotes John Boyd: “People, ideas and machines — in that order” and the fact that he talks about “embedding evidence in the policy process” (though why wouldn’t you?!).

Nonetheless there’s a gap in Cummings’ approach around the more human side of government. There’s an emphasis on hard data science and technology but with little awareness of the value a human centred approach brings to policy or the civil servant experience.

The best projects I’ve worked on around government have combined data crunching with softer insight around behaviour and attitudes gathered through qualitative research. Evidence and empathy.

Government needs to understand the ‘whys’ as well as the ‘whats’ before making decisions that affect lives. Especially when, as Rachael Coldicutt points out, “all datasets are incomplete, all datasets are biased”.

Matt Jukes has written a more detailed analysis of Cummings thinking here.

I’ve written more about the value of a human centred approach on PublicTechnology.net

Enhancing a discovery phase with sacrificial design concepts

Why we use sacrificial concepts as a discussion stimulus in user research

Designing public services in a user-centred way during a crisis

Some of the tools and techniques we use when we run remote discoveries