Are You Preparing for Analytics, or Just Collecting Garbage?

​If you only read one (other) thing today, go read Andrew Oliver’s article on InfoWorld, “With big data, CEOs find garbage in is still garbage out“. ​

Basically, the article reports that a recent KPMG survey of CEOs says that most don’t trust their organization’s analytics capabilities. Not that they don’t see analytics as important, or recognize the value in the practice; they just don’t believe their organization is doing it well. 

Image courtesy of Pixabay

They’re investing in the tools, hiring people with the skills, paying to house and extract and transform and analyze data, and – in many cases – receiving volumes of reports and dashboards from which to base their recommendations. And yet? CEOs are, by and large, skeptical of the value.

It’s a conversation I’ve had no fewer than six times in the last month. It’s a scenario I’ve lived through multiple times in the last five years. And I have to say, the CEOs are probably wise to be skeptical. (And as Andrew Oliver points out, they’re also primarily to blame.) Organizations have suffered for a few years from the Big Data hype cycle, and now, as the cycle moves along its inevitable trajectory-correcting downslide, CEOs are looking at years of budget allocations which have yielded very little.

The problem with the Big Data hype cycle was that it was built off real-world, positive examples of data mining. We heard actual examples of companies that discovered profitable opportunities and/or uncovered massive bottom-line inefficiencies. And we heard examples of ways to apply similar techniques to projects like ours. More than anything, we always knew that data-driven decisions were more effective and efficient than conjecture or guesswork. But gathering large amounts of data is different than planning appropriately for analysis.

The critical aspect left out of the early hype cycle success stories was the readiness of data. I worked with my first enterprise data warehouse in 1999, and witnessed business analysts discover a few opportunities which earned massive profits. But I also witnessed over three years of design, spent mostly identifying the business solutions that the company wanted to solve (many, many, many of them) and figuring out the business rules behind every piece of data, so it could be normalized when it was imported into the warehouse. Yes, solutions and opportunities came quickly, but only because the company had prepared itself for solutions. It had asked three critical questions:
​

• What problem are we trying to solve?
• What are all the ways we can measure that?
• How do we make sure the data is ready for analysis?

There is no tool or technology that will answer these questions for you. Tools and technologies will help you quickly -and accurately – find solutions to your problems (question 1). But no tool can answer these questions for you.