hjThere’s room for improvement in sentiment and text analytics technology: Greater reliability, more accessibility, and increased businesss actionability should be on the agenda.

“There are still tools that are crashing in the middle of a brand crisis or there’s not adequate support or there’s no integration with other tools that need the data,” Chris Boudreaux, global lead, social media and text analytics at Accenture, told an audience gathered at last week’s Sentiment Analysis Symposium. “The industry has to grow up and be more accountable in delivering tools and services.”

Rather than vendors focusing on adding more metrics to their text and sentiment analytics toolkits, he recommended focusing on making the metrics that are already there available in a more reliable and accessible way that ties back to the business. There’s still too great a requirement for too many business scenarios that the people who use these tools require special training or knowledge to gain insights – which themselves continue to suffer from too much variability, he said.

And having those individuals deliver a Powerpoint deck analyzing whether consumers in the social media space feel positive or negative about something is useful only to a limited set of people in limited ways. Those insights into consumers’ sentiment, and that way of presenting them, he notes, don’t necessarily help better inform the call center agent who has to make a decision about how to respond in realtime to a customer calling about an issue. “How can you better inform that?” he asked.

“People want [insights] to be more tied to the decisions they have to make with the business,” he said. Businesses have long leveraged key performance indicators, before the birth of and independently of sentiment analytics, and they could see more concrete benefits from embedding and informing those KPIs with new data sources and sentiment analytics, Boudreaux said. “Give me the KPI I need for my businesses, not just new metrics,” he said.

There was some agreement with that point of view expressed by other speakers at the event. Augie Ray, director of social media strategy at American Express, said that he gets “tired of hearing about every new feature every vendor wants to pitch. I care less about what’s new than what you have done already” – and done successfully. He also explained that AmEx already has a lot invested in Net Promoter Scores (NPS) and Recommend To a Friend (RTF) metrics, with a well-established voice of the customer program – and it’s there that sentiment analytics could potentially be a more effective business tool, rather than assessing 140-character tweets from parties that may or may not be customers.

”Our question is how to go upstream, get closer to the actual transaction itself,” he said. “That’s where our primary interest is, to use [voice and sentiment analytics with] data we already have by recording click-to-chat and email transactions to know how we are doing each and every time we interact with a consumer.”

Market research, Boudreaux said, remains an area where it’s not necessary to expand accessibility beyond the consultants or experts in social and sentiment analytics. “But most people don’t work in market research, and it’s in all these other [areas] where they could utilize all these tools and they don’t,” he said.

IBM is looking into this accessibility issue, as well: Consulting manager Nick Poore in the vendor’s Emerging Technologies group presented at the event the work being done on jStart, a do-it-yourself analytics technology for line-of-business users so that they can “take data, do something with it and get useful output.”