This content is part of the Essential Guide: A CIO's guide to SMAC strategy and governance
Evaluate Weigh the pros and cons of technologies, products and projects you are considering.

Is text analytics the next BI beachhead for CIOs?

Is it time to bring text analytics in-house? A director of analytics at Visa advises CIOs to consider these three questions before making a move.

Text analytics and the CIO don't ordinarily go hand-in-hand. Instead, the emerging field of text analytics, or the mining of text to derive business insights, is typically farmed out to experts. But if the business is serious about investing in tools and bringing the technology in-house, CIOs should be involved, according to Ramkumar Ravichandran, a director of analytics at Visa Inc. And the need to do so could come sooner than CIOs expect.

"We are at an inflection point. Things are changing," Ravichandran said. One notable trend driving interest in the technology is the now-ubiquitous nature of traditional analytics. "The competitive edge from having KPI-driven analytics is slowly going away," Ravichandran said. Companies that want to lead the charge are pushing beyond tradition and embracing more complex analytics capabilities.

It's always a tradeoff between the power of the text analytics versus things like usage and the simplicity of the user interface.
Ramkumar Ravichandrandirector of analytics, Visa Inc.

But, he warned, text analytics is also a niche technology. "There is huge separation between people who use [text analytics] for their whole operation and people who look at it as just an ad hoc thing," he said. In other words, there are masters like LinkedIn, Yelp and Google, where text analysis is core to the business, and then there's everybody else, utilizing third-party service providers for anything more complicated than word clouds.

Ravichandran said the reasons for that gap are manifold: the variety of data sources, the inherent complexities of cleaning up and preparing the data, and the cost of resources. "Today, there has been some investment by vendors to make text analytics technology simpler. But it's always a tradeoff between the power of the text analytics versus things like usage and the simplicity of the user interface," he said.

The biggest reason, however, is that traditional companies haven't created metrics that tie text analytics to revenue or dollar spend. "They're waiting on the sidelines to see if someone comes up with a blueprint," he said. In some cases, the decision to wait until the market matures is a good one. But, for others, while an investment in text analytics technology will bring early adopter trials and tribulations, it could also provide a competitive advantage they're missing out on.

Ramkumar RavichandranRamkumar Ravichandran

Ravichandran, who will be speaking at the Text Analytics Summit West next month, works closely with product management and solution delivery at Visa. He wouldn't speak directly about his work for the financial services company or about the work he did for previous employers PayPal or Move Inc. (the company behind popular real estate websites such as, but he was willing to provide pointers for the CIO based on his experience.

Before taking the text analytics plunge, here are three questions to consider:

1. How significant are the use cases?

Ravichandran said he has used text analytics to address three general use cases. He has used sentiment analysis to find out how customers are interacting with a website, how they're reacting to it, what they like and what their big pain points are. He has done analytical market research on customer surveys for product development teams to give them a better sense of a customer's overall experience. And he has done an analysis of a site's search queries to find out how effective a website's search function is for users.

2. Does your text analytics strategy encompass more than social media?

Ravichandran finds social media channels to be too noisy, a primary reason, in his view, why the business hasn't invested in tools and instead relies on third-party service providers. "Everyone says social media is the next thing, but there are limitations and challenges," he said. That doesn't mean businesses shouldn't be listening to or interacting with their customers via social media, but digging into social media data might be too much work for too little reward -- at least right now.

Rather than focus on dissecting tweets or Facebook messages, one of the trends Ravichandran is paying attention to is online surveys. "People have already realized social media, although exciting, is not the best solution," he said. "So they're running their own surveys, partnering with ForeSee and [Omniture], to come up with their own insights." ForeSee and Omniture are third-party analytics service providers.

3. What were the results of the pick-prove-sell process?

Before going too far down the text analytics rabbit hole, CIOs should encourage the business to utilize the "pick-prove-sell" method. The business should select a problem and perform a proof of concept that can help determine the costs and benefits of the undertaking. In some cases, the right decision will be to sit on the sidelines rather than plunge ahead, and the proof of concept will help sift out challenges and revenue benefits, he said. Ravichandran warned that the problem selected shouldn't be just any problem, but something people can relate to, something that can be sold to sponsors. "This comes from my own personal experience when I knew text analytics was the right way to go, but I couldn't convince the stakeholders," he said. "I did not pick the right question."

Let us know what you think of the story; email Nicole Laskowski, senior news writer, or find her on Twitter @TT_Nicole.

Next Steps

Read how Nordstrom's Matt Fisher discovered a shipping problem by analyzing the unstructured data of customer reviews.

Dig Deeper on Enterprise business intelligence software and big data

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

I think it's likely that you need to pair text analytics with other analytics tools to come away with something valuable - especially if you're trying to analyze social media. Text analysis may give you a starting point toward understanding group sentiment, but you'll need to do further digging to come away with something useful. Similarly, understanding search terms is just the beginning; you'll have to match that with other analysis about your own topical strengths, user research and more.
Ben - absolutely right. Text Analytics has to be used in conjunction with other methods. In my talk, one of my recommendations would be to first verify then trust for multiple reasons (e.g., small sample sizes or mistakes with crowd sourced/manual tagging). But Text Analytics also has a tendency to throws "aha" insights, never before seen or realized. Hence verifying with functional experts, standard reports of established KPIs or analysis of proxies is required to be able to trust the insights.
Thanks for the followup Ramkumar. There's something to be said for 'aha' insights - as long as you don't just take them at face value.