New data tools are introduced almost daily, existing tools have shot to the top, fallen back, then resurrected themselves as established players. Conferences and articles on big data, data analytics, data scientists, and CDOs are never ending.
Instead of focusing almost exclusively on new technology and new roles, there needs to be more emphasis on how you actually deliver high-quality, accessible, and secure data to the people who need it.
Why the focus?
The business benefits gained from strategic information are eroded when your analysts and data scientists have to be their own programme management, data governance, data quality, and IT department rolled into one.
That isn’t why you created their roles – it’s a huge waste of their time which could be much better utilised.
At some organizations, analysts may spend as much as 80 percent of their time preparing the data, leaving just 20 percent for conducting actual analysis.
Booz Allen Hamilton
What’s required is a plan and architecture to manage and deliver data. Valuable, well defined, consistent data with transparent definitions that can be used ‘off the shelf’ by your analysts.
Here are the key steps I think need to be considered.
Start with a high level view of the solution
Information management is a business process, not a technology you buy. Starting with business requirements, you’d ideally stand up each piece of the information management chain sequentially from left to right to build a sustainable foundation to achieve the best results. However, the reality is that because of tight timeframes, this approach is rarely feasible on day one when programmes are launched.
Below is a comparison between the ideal path and the path most taken!
Be pragmatic!
It’s ok – you don’t have to have these information management domains enabled at an enterprise level right out of the gate.
But they are all absolutely required at a basic functional standard if you want a scalable and sustainable analytics capability longer term. I’m not advocating short cuts (which usually take at least twice as long in the end) but a carefully planned, long term architecture implemented in manageable phases.
And then?
Collect data!
Lots of it. You’ve got the required processes and infrastructure to deliver and you can begin to turn raw data into actionable information.
Sit down with your key stakeholders and come up with a list of what the most valuable and reused data is to get the best business benefits. The goal of this approach is to remove as much ‘donkey work’ as possible for the megamind data scientists.
The lion’s share of strategic analysis takes place within familiar concepts like geography, time, products, market/industry, currency, and counterparties like customer, competitor, and restricted parties. There will obviously always be other sources to consider, but you can and should focus on what is eating into the time of your analysts. Start with the core of that and build on it over time.
In closing
The worst thing you can do is to do nothing. The second worst thing you can do is to rush headlong into a technology and roles first approach without a data management plan. The best thing to do is define your strategy and architecture for enabling analytics through high quality data.
I’ve waited over a quarter of a century to find an appropriate use for the parable in the film Colors starring Sean Penn as a young cop and Robert Duvall as his wise mentor. I think this is finally it.
“There’s two bulls standing on top of a mountain. The younger one says to the older one: “Hey pop, let’s say we run down there and kiss one of them cows”. The older one says: “No son. Let’s walk down and kiss ‘em all.”
Don’t make the mistake of the younger bull. Rally your analytics stakeholders and take the time to sit down with an information architect to make the blueprints for high value, sustainable information management. Then walk down to this new opportunity and get your just deserts.
Let me know what your challenges are, I’d love to help.