Where’s the ROI in Big Data?
What is the return on investment (ROI) for big data analytics initiatives? The question may seem simple, but in fact it’s devilishly difficult to answer. Nevertheless, Teradata recently set out to answer that question, and to quantify the value that organizations are finding with big data.
In the study, titled “Betting on Big Data,” Teradata sliced and diced the ROI question from several angles. The analytics giant partnered with Forbes Insights and McKinsey in preparation for the study, which involved 316 senior data and IT decision-makers at companies with average revenue of $500 million.
If the survey had to be summarized in one sentence, it would probably look like this: About two-thirds of companies are reporting solid results from their investments, while about one-third are still searching for a return.
Teradata’s Chris Twogood is mostly happy with the numbers. “I think the big essence of the study is that people are getting value from their investments but it’s hard,” the vice president of product and services marketing for Teradata tells Datanami. “It’s not just, ‘Lets go buy some technology deploy it and it’s great.’ It’s about being able to build data into the DNA of the company, which means you have to deal with people and process and technology, and build a culture to adopt it.”
Finding the Value
Was Twogood fazed that big data improved things by just 1 to 3 percent on average and didn’t show significantly better results? Not really.
“I was pleased with it, especially with the size of the companies,” Twogood says. “It’s asking a lot to take a specific initiative and attribute [value to it]. How do you fundamentally associate an ROI back to an initiative? Depending on the size of these companies, if you can save single digits in your operational efficiency or you can grow revenue without introducing new products or channels” then you’re doing good.
“I think it will continue to grow,” Twogood continues. “I don’t think we’ll hear companies say ‘Big data analytics alone grew my company by 15 or 20 percent.’ I think that’s crazy for a single initiative. But as this evolves, you’ll absolutely see it grow. “
Measuring ROI is notoriously difficult to do in the IT world, but it can be especially hard to do in big data analytics. Companies are spending a lot of time and money on big data initiatives these days, and they’re under the gun to prove the value. Without hard numbers, it’s doubtful that CEOs will keep investing in big data on faith alone.
In Teradata’s study, only 37 percent of respondents could quantify the business case for big data analytics, while 47 percent could not and 9 percent reported “no clear vision.” There’s a lot of improvement to be made in this area, Twogood says.
“I would obviously rather see more quantified business cases,” he says. “A lot of early big data stuff was ‘Go try it, do some sandboxes, try some theories.’ It’s been different than more traditional information management initiatives.”
More Case Studies, Please
Claudia Imhoff, a respected analytics expert and CEO of Intelligent Solutions, says quantifying the business value is big data’s biggest problem. “Part of the problem with big data is that we have become so enamored with the technology, we’ve forgotten what business problems we’re trying to solve with it,” she is quoted in the study as saying.
Without proving the value, big data practitioners risk the ire of those holding the purse strings. “The industry is paying far too much attention to the big data idea without asking the appropriate questions, like, ‘What’s the ROI for this initiative, and how do I measure it?’” she says in the report. “We still need good solid case studies of companies that determined, ‘This was the business problem we needed to solve, here’s how we solved it using big data, and here’s the benefit.’”
Jack Phillips, the CEO of the International Institute for Analytics, agrees that calculating the ROI and proving the value of big data investments should be a priority. But he adds that the nature of big data activities often make that hard to do.
“It’s a murky science still. It’s so hard to know,” Phillips tells Datanami. “If you break ROI into components, it’s actually hard to calculate the ‘i.’ But even more challenging is to work with the business enough to isolate the contribution of analytical activity to whatever the improvement is.”
Say a retailer enjoys a 20 percent boost in revenue after the algorithms told it to raise prices during a holiday. Yes there was an increase in revenue, but there was also an increase in volume. Getting a clear read on the increase due solely to the algorithm and separating out the other factors is difficult to do.
A Seat at the Table
The good news for the analytics team is that the impacts are positive and they’ve got a seat at the table. “They’re in the conversation,” Phillips says. “It’s a precision question. We know it’s positive, but now everybody is saying, how positive? The science will mature and we’ll get better.”
Twogood agrees. “I don’t see investment slowing down at all in this space,” he says. “I think the ROI stuff will get tighter and tighter as we move forward. The more [positive results] they get, the more they’ll want to measure it tightly and get more results, and that will accelerate the investments.”
The big data analytics space is maturing rapidly on a number of fronts, and measuring ROI and value is definitely one of them. In many ways, big data analytics will take on traits of traditional data warehousing, including ROI measurements.
“Companies will start to ask themselves ‘How do I apply the kind of discipline they had in the data warehousing space into the big data space,'” he says. “The last couple years has been, ‘Let me go try it, let me see if there’s value in data exhaust and how do I get movement and insight that I wasn’t able to get before.’ We have an ROI model and there are others that have models. We’ll start to see more and more of the models applied to big data space.”