Big Data Grows Up

Over the past decades, analytics processes have been developed and deployed in a custom, one-off, artisanal fashion. This makes sense when high-value problems are being addressed, but not when it comes to affordably incorporating big data into the myriad small decisions a business makes every day. For big data and analytics to reach their full potential, they must be made operational. And making that leap will be akin to the industrial revolution a few centuries ago.

From Custom To Affordable And Scalable
The industrial revolution took manufacturing from a labor intensive, fully custom, non-scalable endeavor to one that can produce affordable, quality items at massive scale. We similarly must extend big data and analytics far beyond what is being achieved today—they must be embedded and deployed much more deeply and broadly in the enterprise. Yes, we’ll have to give up some of the customization and perfection that we’re used to, but that’s a trade-off worth making.


Just as we still buy custom pottery as a centerpiece, we’ll still have custom analytics processes in cases where they’re warranted. At the same time, we’ll also be able to affordably address many other issues with analytics that are:

Embedded and Automated
A fully embedded, automated analytics process can run and make decisions whenever required until somebody explicitly turns it off. This can be thousands or millions of times per day in cases such as optimizing the offers displayed on each Web page.
Prescriptive

We are now able to move into the realm of prescriptive analytics. This not only predicts what might happen, it also prescribes the actions needed to make or prevent the event from happening. Instead of simply predicting that an engine failure is imminent, operational analytics will determine what should happen to avoid that failure. This prescriptive component is necessary since human intervention can’t be required.

Executed in ‘Decision Time’
Not every analysis needs to happen in real time, so operational analytics can be configured to be executed within the timeframe required for the decision. For example, if a business posts refunds at the end of each business day, then any analytics required to check for fraud simply must be completed before day’s end.

Designed to Take Action
The ability to cause a recommended action to happen is what most distinguishes operational analytics from traditional methods. This can include extending an offer or issuing a warning as a result of the analytics. The process is no longer a passive one that supports a decision maker taking action, but rather the analytics process becomes the decision maker taking the action.

Experience the Impact 
Making analytics operational does not replace traditional skills and approaches. It takes them further. Qualified people will still need to use traditional methods to identify the business need, build and test the analytic process, and validate that it works as expected before turning the process on and making it operational. The analytics revolution is an evolution of analytics—taking it to another level—not something totally new.

While examples of operational analytics encompassing big data already abound around us, most people aren’t even aware of how it all impacts their daily lives. And that’s okay. The point isn’t for people to be aware of and appreciate the analytics, but to have them notice the improvement in their lives. It might be simple analytics that automatically change a home’s environmental settings to balance comfort and cost, sophisticated processes that reroute travelers following an adverse event, or the emerging self-driving cars that constantly analyze a wide range of data to decide how to get you to your destination safely. None of these would be possible without data and analytics.

What is your organization doing to take it to the next level? Have you joined the analytics revolution yet? If not, what are you waiting for? The future is coming fast and you want to be prepared.

Credit: Forbes