"Big Data" from "Little Data"
"Big Data" has been the promise as the "Next Big Thing" in technology. Most companies I know, struggle with how to create real impact from Big Data as this McKinsey Insights article discusses. Many are travelling a journey to realizing that impact. In my experience, this is the rare time when technology may have to lead. Business leaders understand the promise of "Big Data" but rarely know where to start. I suggest building support organically by strategically starting small, before launching into big organizational change programs.
I believe business saavy IT leaders need to take the initiative to make this impact a reality. A good first step is identifying a great subset of "Small Data" that is widely used and important to business leaders. For example, key business indicators or critical business metrics that are widely used and important in the organization with the following characteristics:
The data is "the truth" -- although a lot of people might debate the accuracy of the data, inside your organization, this data is widely regarded as "the truth". This "truth" might be referenced frequently in business meetings, especially performance related meetings; individual performance, group performance, division performance or company performance. If the data you are looking at is that kind of data, that is a great target for your "little, big data" trial. Some examples are as simple as sales data, utilization data, unit production data, quality data, pricing data.
"The truth" is frequently manipulated or normalized -- another good indicator of a good target for your "big data" entry. Since "the truth" cannot be challenged, many business leaders are called to explain indicators that come out of "the truth". Lots of times, that means normalizing the data. Although this is very helpful to the business leader trying to explain why the metrics look the way they do, it is frequently frustrating for C level executives because after everyone gets done normalizing the data the clarity from "the truth" gets lost. Some examples include headcount; a Human Resources executive might count Full Time Equivalents one way, while Operations executives measure them a different way.
My recommendation to saavy IT leaders is to find "big data" tools that will allow business leaders to analyze and manipulate these critical data elements. The best tools will be flexible enough for manipulation on any device; laptops, tablets or smartphones. Once a tool has been identified that fits in your architecture, then you need to sponsor an initiative to take that very small, well defined set of data and create a normalized data set for the tool to use. Throughout this process you should work with business leaders across constituencies to understand how they typically want to use, analyze and see the data and make sure your tool will accommodate those use cases and make sure the data is normalized in a way that allows those kinds of analyses. Once you have the tool in place and the data ready, let the executives start playing with the analysis....and I do mean the executives (see article exhibit on "Targeted Analytic Solutions" and "Self Service Tools"). If the tool isn't simple enough for them to do their own analyses, drill downs and compares, you have the wrong tool. At the risk of getting too technical, I have had a lot of success with Oracle's OBIEE tool.
The business benefit will come when C level executives are comfortable that everyone is using the same "truth" to explain their analysis of the data. I assert that once you plant a seed like this, the "little data" will become "big data" and as the scale of the data and analysis grows, the impact will grow.
I guess my commentary to proceed the introduction of this great McKinsey article has gone a little longer than I expected. Here is the link to the article.
http://www.mckinsey.com/insights/business_technology/getting_big_impact_from_big_data
And another...
http://www.mckinsey.com/insights/business_technology/big_data_whats_your_plan
