Big Data Technology—the Rebel without a Cause
Gartner cited the #1 challenge in Big Data as “Determining how to get value from big data.”
Did I read that right? And by no a small margin! How to get value from big data? Shouldn’t it be shocking that such a fundamental question persists while businesses are spending billions on data analytics capabilities and infrastructure? If ever there was a solution looking for a problem (or a rebel without a cause), this is it. How can the most elusive part of big data be “how to get value?” Shouldn’t someone have thought of this “before?”
We all talk about the opportunities to enhance customer experiences, uncover new monetization opportunities, and increase operational efficiencies. But… where’s the beef??? Well, here is my take on it.
The graphic below is over 1 year old – I know that. But the majority of organizations that I speak with are in this same boat today – how to get value? Well, we know that IT has been getting pressure to develop capabilities to support big data initiatives. So IT is investing in the capabilities they think they need. And there are plenty of vendors that are more than happy to sell them the kitchen sink. In fact, these purchasing decisions are relatively easy—the technology works and there are many choices. Again, see below in the red box (the infrastructure and architecture are not significant challenges for customers).
However, there are many examples of wonderful successes with big data and analytics and we read about those in the news, and we experience them in our everyday lives (Netflix, Uber, AirBnB, Amazon, Tesla, etc. etc.) But for the thousands of businesses you don’t hear about, why are they finding it easier to spot a snow leopard in the wild, than figure out how to get value from big data? Technology is not the problem. Here are a few things to keep in mind that will help you learn how to integrate data and analytics into your business models:
1) Begin with an end in mind. Thank you Stephen Covey. In addition to the chart above, Gartner recently reported that many big data projects are not making it to production – only 15 percent of business they surveyed reported deploying their big data project into production. What we tend to see is that these projects stall out as science experiments because nobody bothered to determine the potential ROI of the initiative, and then priorities change. In our consulting organization, we offer a Big Data Vision Workshop to help customers identify and prioritize the analytics use cases that have the best combination of business benefits and implementation feasibility. It’s wonderful offering and I’m not aware of another like it. But we don’t stop there. To help validate the potential ROI of an analytics use case, we have a Proof of Value service where we use real customer data and show them, in their own IT environment, how to surface the insights and how those insights can be consumed from an end-user standpoint. After all, insights need to be actionable, or there is no value.
2) Big data solutions are not compilations of IT infrastructure. There are many IT infrastructure vendors that are selling infrastructure but calling it a big data “solution.” And I say… wrong. They may be selling a very robust platform with every possible capability, but it’s not a solution until it is solving a problem. It’s a rebel without a cause. Don’t get me wrong, the hardware and software are required, but that’s not the top challenge. Fortunately, I work for a company that understands that the customer is running a business, with business challenges and business opportunities.
3) Think Like a data scientist. Our Dell EMC Services CTO, and accidental mentor of mine, Bill Schmarzo, has helped many of business leaders to learn to think like a data scientist, and it’s not as hard as it sounds. The process is fairly linear: consider a top-level strategic initiative. Consider the business stakeholders. Consider the decisions they are trying to make in support of the initiative. Consider the questions they must answer to make those decisions. Consider the data sets that are relevant to answering those decisions. There is more to it than that, but if you can follow those basics, you can lead your data science team to focus on the data and analytics scenarios that are going to help you achieve your business objectives. Of course you’ll have to examine the implementation feasibility as well, and that’s why the list of stakeholders must include people from the business and IT.
I could go on, but since a picture is worth a thousand words, I put this infographic together to help customers understand our approach to success with big data. Enjoy!