Accelerating Big Data Time-to-Value

Big Data can become Big Impact Data for your business. Analysts are watching the evolution of Big Data from trendy to valuable insight.

I like this technology maturity curve from Gartner. In 2013, Gartner places Big Data close to the Peak of Inflated Expectations, which means that it’s 3-5 years away from the Plateau of Productivity.

That’s a long time!

What if you could skip the Disillusionment phase and start making money from Big Data projects a lot sooner?

I asked the advice of Chris Ward, Brian Vaughan and James Bigger, Principal Consultants with World Wide Technologies ( and this is how their customers have done just that:

Step 1: Have a clearly defined business use case.

Work with the line of business executives and together precisely define a Big Data use case that is core to the business and has the potential to create transformative value through leveraging multiple data sources and predictive analytics. Some examples:

  • A retailer that leverages many data sources (marketing offer/response, historical transactions, web-site browsing, 3rd party profiles and more) and recommendation analytics to improve the personalization of marketing campaigns, coupons, and offers to compete better with competitors.
  • A railroad that uses sensor data, microphones and ultrasound to capture data about their engines to identify equipment at risk for failure
  • An investment bank that leverages the portfolios of their more than 30,000 FAs to create apples to apples performance comparisons and generate product recommendations across this network 

Step 2: Translate the business problem into a math problem and leverage as much relevant data as possible

The two critical components to building highly predictive models are:

  • Attack the business problem with multiple analytical techniques; often it will be an “ensemble” pulling the best of many models together that delivers the breakthrough in predictive “lift”
  • Build a “data ecosystem” that leverages and links as many new, relevant data sources as possible; remember there are usually treasure troves of data inside the 4 walls (both structured and unstructured) that have never been properly brought to bear on business problems or opportunities

Step 3: Think strategically from the beginning

This is a new area of investment for your company, so it makes sense to start small; but starting small doesn’t mean starting cheap. To make the right technology decisions you will need to understand what will be required to operate the solution at scale.  Calculate your ROI based on your TCO, not just based on acquisition cost. This will provide a clearer picture for your CIO and CFO and will get you funding for the next business use case.

Step 4: Test often

Test your initial assumptions with a Proof of Concept as close to production as possible. Adopt an agile development methodology so you can discover mistakes early and adjust course.

Step 5: Communicate

Over invest in committed resources on the first case study to increase the odds of success; then broadcast the results widely and loudly through all channels.

WWT has built a Flexpod Select Reference Architecture in their Advanced Technology Center (ATC) to highlight core features of NetApp’s products in a Big Data solution.  Specifically, WWT will plan to leverage the ATC to demonstrate key points of differentiation in the NetApp solution – separation of compute and storage, enterprise grade reliability, market leading data replication factors that drive compelling TCO at scale, etc. 

By offering customers an ATC facility that provides practical advice and comparisons on the hardware and software components of a Big Data stack, demos that highlight how technology and analytics drive value and the ability for customers to load data and run POC’s, WWT can really help with the strategic decisions of starting small but thinking big.