Guest post by Graham Brierton for Big Data Republic
As we move from a busy 2013 for big data and into what is shaping up to be an even bigger year — with IDC telling us that the big data technology and services market will grow at a massive 27% compound annual growth rate (CAGR) in the next five years, hitting $32.4 billion by 2017 — it seems timely to remind ourselves that technology won’t do this by itself.
In other words, having gone to the conferences, won the budget, and completed the training to get your big data team ready, you are still not quite home and dry, despite all the great groundwork.
Why? Because the main lesson from the history of information and project management is this: IT projects don’t fail because the basic technology doesn’t work, they fail because the end users couldn’t use the system (the functionality was inappropriate), or because the user experience wasn’t tuned to their needs.
You can have the most sophisticated system in the world, but if the users don’t buy in to it you’ve simply wasted your time and the business’s money
The Winds of Change?
Are things changing? Possibly. In our era of software-as-a-service and the cloud, enterprises are starting to learn a lot from the way consumers use the web. These savvy customers are asking, “Why can’t the communications software at work be as easy to use as the free stuff I use at home? Surely I could do this with my Apple, not your clunky green screen ‘application’?”
At the same time, vendors now have the ability to gather data about users’ experience at scale. This means they can flex and change the application much quicker, based on this feedback, which is essential to delivering easy-to-use applications.
That means that how we bring data in, generate queries, and ask questions of the data are increasingly things we have to think about in terms of the suitability for each level of that data and the type of user engaging with it. I say “suitability”: if the data helps the user do a particular job. To decide if it will, we need to ask another key question: Have we involved those users in any validation exercise here? That’s surely the only way to ascertain what other tools, or data, or personal interactions are needed to get the maximum value out of this big data?
The point is this: Although the tech may have changed, that underlying dynamic, that it’s people who will consume your data, and it is people who will try to act on it, is more true than ever, even in the big data era.
Some of the users love numbers, columns, and standard deviations; others like visual to be the dominant theme, with aids to make it easy to collaborate with others on the report or on the decision before it has to be made. To be useful, the data ought to be able to be interacted with in different presentation styles for different role types, i.e., working from queries for power users, in a “click and drag” style for creators or in a visual touch fashion by information consumers and so on. The point is that data should always be available to the user within the most appropriate workflow for them.
Getting the balance right
You need to give the team what they want, the way they want it. So as you draw up your ambitious 2014 big data plans, I strongly advise you to take some time to make sure you really understand how the end users of the system you are building want to work with it.
If you don’t, some depressing IT history may repeat itself, and none of us wants that.
Here’s an interesting final thought: While the data may need to be presented in a format that users will find intuitive, the data itself may actually deliver results that at first appear counter-intuitive. For example, SAP’s Sameer Patel believes that big data and algorithms often may offer an unexpected set of suggestions, as the numbers do not reflect our embedded learning and experience about what “the right next thing to do” should be.
In other words, making things easy for users sometimes, if we use big data, may result in some difficult answers that we may need to work with our users to best understand.
Read more on Big Data Republic
More blogs from Graham Brierton
+44(0) 20 358 243 60