Scrollen

Einblicke und News /

03 September 2018

How to sort out Data Analytics and AI: Data Strategy (Part 1)

David Blackwell

David Blackwell
Partner | Data, analytics and AI | London

“Data” and “Strategy”, two of my favourite words nestled together in one special phrase.  They belong with each other just like Milk and Cookies, Beardsley and Lineker, Middle Aged Men and carbon race bikes. 

What is even better is that many organisations are now waking up to the fact that they need a data strategy.  Having liberally splattered investor presentations with share price-plumping CTRL+V phrases like “We treat data as an asset” and “Data is the new oil”, a lot of businesses have failed to live up to their own data hype because for all the grand intentions, they haven’t backed it up with a strategy.  Yes, organisations have spent money on warehouses, lakes and data scientists – but in many cases they haven’t properly thought through how data will drive value for their business and invested accordingly. 
There are a lot of white data lake elephants and bored/frustrated data scientists out there…
 
So what is a data strategy?  In short, it is an expression of how data will be used to deliver the business strategy. 
 
The first step on this journey is answering all the “W” questions (I will come onto the “how” in a subsequent post):  What data will we need? When will we need it?  Where will it be needed?  And most importantly, why will we need it?
 
Now, when seeking to answer these questions with different people across an organisation, it can be a case of “feast or famine”.  Some will have a very clear idea about what data they need (and will likely have spent years getting frustrated that no-one can deliver it for them, and hacked together a skunkworks version of it already using stickytape and SQL) and some will not have a clue. 
 
To smooth this process out, and ensure we are being exhaustive, there are a couple of simple techniques we deploy…
 
Firstly, we peer at the question through some different lenses: 

  • Offering: What data is needed to support the products and services the business wants to offer?
  • Customer: What data is needed to underpin the various customer journeys?
  • Operational: What data is needed to support the different capabilities or processes within the organisation? 
  • Risk: What data is required to address and mitigate the risks we face?
And when looking through these lenses, we are thinking about what it takes to run the business (e.g. what data is needed to make a process actually work), but also what insights we might want in order to improve the business, and what data we might want at our disposal to enable the automation of tasks through AI. 
 
These lenses clearly overlap and are not mutually exclusive, but they do enable us to come close to being exhaustive when it comes to pulling together a view on the data an organisation needs.
 
Secondly, we bring some ideas to the discussion.  Starting with a blank sheet of paper can be daunting, the infinite expanse of a question like “what data do you need” can feel starting to write a novel or trying to paint the extremities of the universe with finger paints.  So it makes sense to bring a few examples of how data could, for example, enable customer on-boarding, improve customer service, make a field force more efficient or reduce employee churn. 
 
Time’s up for this blog, but in the next edition I’ll start to get into prioritisation, personas and the practicalities which underpin a data strategy.