Discover Performance

Big data expert and author Edd Dumbill explains why being data-driven means never throwing anything away, and tells you how to successfully make the cultural shift that will get there.

Enterprises are increasingly driven by data. Drawing competitive advantage from the rapidly growing body of data that businesses collect is not just a trend, it’s the new baseline.

The resulting data-first philosophy means abandoning many long-held beliefs and best practices for collecting and using information, explains Edd Dumbill, author of the recent ebook, "Planning for Big Data," and program chairman for the Strata Conference at O'Reilly Media. One of the most critical changes an organization must adopt, says Dumbill, is to stop throwing data away.
 
The traditional way of processing raw data is to immediately extract what’s relevant for your analytics systems, such as accounting or inventory, and then view only what provides recognized business intelligence value—sales by product or by branch, for instance. The rest ends up in the virtual trash.
 
In a data-driven organization, Dumbill explains, there is no garbage data. The proper way to handle analytics is to put a view on the data at the point of analysis, not at the point of storage.

"When you store everything, you have all of the data at your disposal and you increase the chance of finding those moments of savings or opportunity," Dumbill says. "Even data that ultimately constitutes an error is valuable to have. If you throw it out, you'll never be able to chase it down and eliminate the problem."

When cultures collide

A startup can embed a data-driven culture from Day 1. But for long-established organizations, the transition will be more difficult. Dumbill offers the following advice to organizations that want to make the transition to a data-driven culture as painless as possible:

• Put curious thinkers in downstream roles and give them the freedom to see what they can find. Entrepreneurial thinking, most often found in business development or product development roles, now needs to be equally valued in data analytics.

• Make sure key managers are receptive to the data discoveries that bubble up from big data analysis, even ones that seem ordinary. All improvements are worth pursuing.

• Reassure employees that data analytics are designed to help them, not embarrass them. Data analysis can cast a very harsh and unforgiving light. Mitigate defensiveness by discussing data goals early and often.

• Make data self-service, for as many employees as possible. By providing a data platform where people can export data and look at it for themselves, organizations drill home the importance of the data-driven cultural shift, and they may reap many new insights in the process.

• Create visualization tools that will give non-statisticians a way to understand the data that's important to them. There are very few people who know how to read data in a statistical way. So if you want to drive the business through data, you need to find ways to bridge that data-literacy gap.

• When possible, provision your data visualization as special purpose mobile apps. Mobile devices such as iPhones and iPads have proven to be very effective delivery methods for special-purpose applications. By working within this trend, you give your apps the best chance of finding strong adoption in the organization.

First steps

Read more on how to overcome the challenges of moving to a data-driven business philosophy in “Big data’s big changes can start small,” posted on the Discover Performance blog.
 
Learn more about Edd Dumbill's free ebook, "Planning for Big Data: A CIO's Handbook to the Changing Data Landscape," at the O'Reilly store. For more on using business analytics that can truly monetize data, visit www.vertica.com. For more on solutions to managing big data and extracting intelligence from unstructured information, visit www.autonomy.com.