justinkendra - Fotolia

Is aggregated data always useful for business?”

Storage is cheap, but that alone cannot justify collecting data forever. In this podcast, IBM's Mike O'Rourke offers ways for dealing with aging data.

Businesses are collecting mountains of data. Over the long haul, does that data increase in value, like a fine wine? Or does its value plummet, like forgotten-about refrigerator leftovers?

"That's a great question," says Mike O'Rourke, IBM's vice president for business analytics. "There are some kinds of data that get more valuable over time." Speaking of wine, O'Rourke says historical data on growth, weather, climate, harvest and other factors may be very valuable when analyzed over multiple decades.

But there are other ways to aggregate data, such as from Twitter, "where you have [a lot] of information. Twitter's keeping all that data, but I may only want a piece of it." In that case, it's necessary to select only the desired data, and then either store it or analyze it.

"You should depend on data providers to make the decisions on how valuable their own data is," O'Rourke says. "For most people trying to develop an application or look at analytics, they want the data and analytics when they need it and then move on. They shouldn't have to store all that data."

Sometimes, it's necessary to keep everything, because there may be a future value in selling that data to third parties. Unfortunately, what may develop future value is not easily predicted with aggregated data.

In some environments, such as retail, "It totally makes sense to aggregate data [as it ages]," O'Rourke says. "That's got to be a part of the overall plan."

If there's one question about aggregated data to which no one seems to have an answer, it's asking what percentage is actually ever accessed. Even a data expert like O'Rourke has no idea.

"My guess is that it's under a percent," he says. "In terms of the data we're pulling for customers and the things they're looking at on their dashboards and not getting any deeper on, it's definitely less than 1%.

Next Steps

How businesses are doubling big data value

Data collection from Windows 10 causes concern

Discussing data access with BYODs


Dig Deeper on Topics Archive