Big Data Strategy: Are You Asking the Right Questions?
by Andrew Donaher
February 24th, 2015
Strategic decisions can often fail if a sufficiently wide perspective isn’t taken, or if enough possible alternatives are not explored. Confirmation bias can lead you to gathering self-serving information and rushing into making a decision. When you have a hammer, everything seems to look like a nail.
The same is true of Big Data.
I love talking about the “what” of Big Data; whether on-prem Hadoop with the whole Zoo, Cisco’s appliance, Cloudera, Pivotal, HortonWorks, AWS, Data Lakes, SAP HANA, ExaData, IBM Watson, MemSQL, etc. – these are all terrific options. And all great stuff to an analytics guy like me. The key, however, is asking the right questions.
When we think of Big Data we traditionally think of internet companies storing pictures, videos, information about our behaviour and social media activity to perform web analytics and real-time offers. Some companies require this capability; others need different capabilities to gain a competitive advantage.
I know of an Oil & Gas/Energy company using Hadoop not for what we traditionally think of as Big Data in terms of size but as a repository for wide variety of data (yes that is one of the big data “V”s) to solve a compute problem for geological analysis. This is a great example of widening one’s perspective and focusing on the business capabilities. The company asked the right business questions and widened their perspective to determine the best way to deliver the capability.
If someone starts with the question “which is the best Big Data solution?” run for the hills. The answer is the one that best suits your needs and profile.
- What business capability do you need?
- What is the value of the problem?
- Is this a one-time issue or something we need to address on an ongoing basis?
- What is the most appropriate cost model? OpEx metered pay-as-you-go access or do you want to spend CapEx dollars?
- What is your skills profile and that of the local market?
- Who are the analysts and end-users?
- Etc.
Big Data is becoming more and more mainstream every day. Vendors are building appliances and solutions that abstract out some of the complexities so people can leverage their existing skills sets to take advantage of opportunities with big data. Informatica is a great example of that with the way they are handling big data integration. Lines of business are now becoming more self-sufficient as traditional IT skills proliferate across organizations and solutions become more and more available. Hadoop and other Big Data solutions will become as important to your ability to get information as your ERP and EDW in the not too distant future. They will all exist together.
This is where IT leaders need to continue to maintain a wide perspective, be responsive and agile. Be prepared to fail-fast, explore options and continue to break-away from the traditional command and control model of IT. There is an insatiable demand for information to make timely business decisions and IT groups need to enable and support access to appropriate data - big, little or otherwise.
Clear success metrics, agile governance and a focus on integrating and provisioning trusted accurate data is the key to driving business value.
comments powered by Disqus
