There is lot of talk at the moment about data analysts or data scientists, but what do you need to be successful in these roles and what type of person do you need to be? The stereotypical view is that we’re ‘a bit nerdy’ and ‘walk around in white coats’, but while that may be true of some, it’s certainly not the norm.
You need a strong foundation in maths, and that’s what I most enjoyed
at school. But like most other kids, I also wanted to be a footballer
or a rock star. And while any data analyst needs to be numerate and
understand data, the best analysts I’ve worked with are those that are
the most creative. By this I mean having a different way of looking at a
problem, which then gives you ideas on how best to solve it. You have
to be sure, at the outset, that the problems you’re solving are the
There are usually so many problems you
could solve, but the key is solving the problems that yield the most
value to the business. If you take supermarket chain Tesco, their
loyalty card delivered by Dunnhumby is built on SAS analytics. You give
vouchers relevant to the customer so they then use the vouchers again
and again. If it means you reduce customer churn by 5 per cent, you can
multiply the average shop per week by the number of new customers,
multiply that by the number of weeks in a year and give the customer
something in millions of pounds which equates to the additional annual
revenue. Giving an answer in pure business terms like that is what
executives will value most. They will be less interested in the maths
that got you there.
In terms of qualifications, we’re not all pure mathematicians either.
My team comprises people qualified in applied statistics, econometrics
and even social sciences. The key is they are all turned on by data,
what you can do with data and how you can solve problems. But do not
expect business people to be turned on by data. We need to present
findings back to them in a format that is right for the person we are
speaking to. If it’s the CMO of a Telco, don’t give them tables and
tables of data. Use visual analytics to display the findings in pictures or easy-to-follow graphs.
The other way we can get our message across is through simple
story-telling. There’s the example of the father who rings up a retailer
and complains they’ve been sending his 16-year-old daughter vouchers
for baby products. He later has to call back and apologise as it turns
out his daughter is in fact pregnant but he didn’t know. This shows the
power of analytics – they understood from her profile, her browsing
behaviour and all the other data available that she must be pregnant, even though her own father was not aware she was.
Another example is a retailer we’d worked with that had done their
forecasting by gut feel for many years. We then used a forecasting
engine and effectively turned it on without telling them. We were able
to show them how wrong they were. They changed overnight from being very
much against the solution to being a big supporter of it.
However, it’s important to tread carefully. A barrier can be the fact
it’s a bit personal if you’re challenging someone’s experience,
sometimes gained over many years in that particular sector. The solution
is very often a mixture of this business experience (and asking the
right questions) combined with an analytical approach.
As for whether most of us are a bit like Mr Spock, well that perhaps
used to be the case. I think that’s changed as we’re much more used to
data, and many current executives spent even their formative years with
data all around them. Many MBA courses now have some element of data or
data analysis involved. Then there’s current graduates born in the early
1990s and therefore born with the Internet. I think we’ve seen
analytics emerge from the backroom into the boardroom.
Learn more about the human side of big data and high-performance analytics in this research report about data scientists by Tom Davenport.