Revealed: Actual ROI on big data investments today


As of this writing, the new Wikibon report wasn't released yet (although it will be by the time this report is published) but analyst Jeff Kelly gave me a sneak peak and his take on the results ahead of time. Some of the findings reiterated what many of us already knew but one thing in particular caught my attention. "Almost half, 46 percent, of companies are currently only seeing about 55 cents return on every dollar in big data investments," said Kelly. "But they should see around $3 to $4 on every dollar spent in the next three to five years."

The three most common reasons for the lag in ROI right now, he said, are a lack of skilled big data practitioners; the technology is too raw and difficult to use; and, the lack of a compelling business use cases.

Kelly says that while over 90 percent believe that big data is a critical element to the competitive modern enterprise and nearly 60 percent believe that big data is the new source of competitive advantage, most have yet to put big data to the test in an actual business use case. So, what is it that they are doing with big data now?

"They are experimenting with big data but not tying it directly to specific projects with measurable outcomes," says Kelly.

Ah, once again we are reminded that big data is still in its infancy and first adopters have to tinker and tweak the technology to figure out how best to use it, and why to use it, before they can actually do much with it.

So once the tinkering is done, how can companies push forward and realize full ROI?

Actually there are several ways to proceed, but here are the top two according to Kelly:

  1. Tie investments to specific projects and outcomes. "Start small with a highly targeted project related to your core business," he said. "Build benchmarks and get executive backing before you take on a bigger project."
  2. Hire a specialized services firm to fill the skill shortage. "The tools will eventually get easier to use and so will resolve the shortage of talent in the long term," he said. "But for now it's better to hire a specialized services firm rather than attempt to hire talent that is hard to obtain now."

Think Big Analytics is one such firm. It has around 60 employees including a bevy of data engineers and data scientists. I had a chat with Scott Rose, who has the rather intriguing but odd title of client partner at Think Big, to see what he's seeing from that end of things.

"We're not seeing clients confused about what they want to do with big data or how to use the technology," said Rose. "The challenge our clients are struggling with is where and how to get started."

Think Big Analytics sends small teams to help out on everything from advising to implementing. I suspect their competitors do much the same. But using a specialized services firm, as Kelly suggested, has more than one benefit. For example, maintaining a current big data investment is often overlooked even though IT people are well aware that maintenance and optimization of any tech is very important. A services firm can tend to such details for you. Case in point: Think Big Analytics' recent announcement (yesterday for you dear reader, but tomorrow from where I'm sitting) of its cluster optimization program which includes a review of clusters, applications, configuration and operations.

Given the current talent shortage, Kelly is probably right. A good specialized services firm could prove a valuable patch until the tools get easier to use and less specialized talent is needed. Just for the record, I am not recommending Think Big Analytics, as I never recommend or endorse any given vendor. However, I do find vendor input on occasion to be truly helpful. Certainly it is good to know what they see in the course of their daily work.

Related Articles:
The future of big data: cognitive computing
Big data strategies: fixed mountains vs. shifting dunes
From the frontlines of the data storage vs in-memory war