The Many Styles Of Social Media Analytics
Whether provided as self-service or with the backing of consultants, for marketing, CRM, or operations, social media analytics are snowballing in importance.
February 17, 2012
10 Cool Social Media Monitoring Tools
Slideshow: 10 Cool Social Media Monitoring Tools (click image for larger view and for slideshow)
Why is social media analytics important? Let me count the ways. Actually, I'm not sure how to count them.
Last week, I presented on "5 Styles of Social Media Analytics" at a Virtual Enterprise 2.0 event, along with some other great speakers. The "5 Styles" label was supposed to be a working title that I would go back and update once I figured out what the right number was, but then I got distracted. Smart editors have also told me I should learn to love lists of 5 things or 10 things because they're traffic magnets. In this case, I still haven't figured out the right number, but it's more than 5.
I knew from my reporting that, although there are some common base technologies and techniques applied across all of social media analytics, there is great variation in the simplicity or sophistication of the options on the market, ranging from basic keyword filtering to advanced text mining and natural language processing. They can provide self-service or white-glove service. The basic principle of extracting trends or knowledge from social streams is being applied in many different ways, sometimes with industry-specific variations.
[ Why should Facebook matter to the enterprise? Read Facebook: The Database Of Wealth And Power. ]
Many of the social media listening platforms got their start providing public relations and marketing departments with next-generation "clipping services," monitoring blogs and social media for brand and product mentions, much as they had previously watched for mentions in the traditional media.
Market research applications treat social media as providing a variation on the kind of intelligence obtained from polls or focus groups, looking for marketing opportunities, threats to their reputation, and sometimes new product ideas. Customer support teams favor monitoring tools that provide some sort of support for "engagement," the ability to respond to a complaint or question aired in social media rather than letting it go unanswered. I'm intrigued by the possibilities for feeding issues identified through social media monitoring back into operations, as in my case study on Social Analytics Cracks Case Of The Jalapeno Cocktail.
"For organizations that talk about social media as it if were a single capability, this view is often a rude shock," Gary Angel, president of the consulting firm Semphonic, writes in the process of giving his own breakdown of the types of social media measurement and analytics. He comes up with six categories, each of which requires a different set of metrics and base technologies. It looks something like this:
-- Customer Support -> Operational Metrics
-- Public Relations -> Influence Categorization & Tracking
-- Campaigns -> Listening & Web Analytics
-- Communities -> Engagement & Attrition
-- Social CRM -> Customer Data Integration
-- Products & Customer Research -> Advanced Natural Language Processing
I also consulted with Forrester Research analyst Zach Hofer-Shall, who helped me out on a panel about big data and social analytics at Enterprise 2.0 Boston last year, and Marshall Sponder (aka WebMetricsGuru), the author of the book Social Media Analytics (McGraw-Hill, 2011).Hofer-Shall said he sees a "spectrum" of social media analytics products, with one axis being the contrast between self-service tools and full-service consultants. "Almost everyone is somewhere in between," he said.
Radian6, now a division of Salesforce.com, provides tools for running queries and creating monitoring dashboards. Radian6 is not necessarily the best tool in any of the categories it has entered, but it covers the broadest range of the basic scenarios, Hofer-Shall said.
At the other end of the spectrum, some of the more sophisticated market research services like Converseon include a heavy dose of consulting, where clients may be handed a bound or PDF-formatted report, including a human analyst's interpretation, on top of whatever the software may divine. The consulting operations may also provide a dashboard, but a customized one rather than a do-it-yourself kit.
Whether consulting services or software as a service, social media analytics are typically procured as services, often by marketing departments, and with little or no involvement from the enterprise technology organization. However, there are some emerging examples of how social data and enterprise data can be combined for more effective results, such as HP's correlation of social signals with sales and service metrics. CIOs who have not yet paid much attention to social media analytics, because of its outsourced nature, ought to at least be thinking through the possibilities.
"Is IT really involved? Definitely not. But they're starting to be involved a little bit," Hofer-Shall said.
Sponder, who previously was best known as an authority on Web metrics and search engine optimization, said he is no longer as interested in search, now that he sees the potential of social media analytics. "The Web is about what's being said all over the place, not necessarily inbound traffic to your website," he said.
However, the tools for social media analytics still have some growing up to do, he said. Many of them were designed for public relations and marketing people who merely wanted to extract a listing of brand and product mentions that they could scan manually.
"The marketing, PR, or communications user tends to want to look at things in an exploratory way," he said. "It's good to find out what people are saying out there, but it can't be scaled and can't be automated" when you take that approach, he said. "It's nice to know about buzz, but there is not anything you can do about it."
More sophisticated approaches to the problem apply natural language processing techniques, seeking to make computers understand the content of an article, or a tweet, or a comment, not just index it. Sentiment analysis technologies that score content as positive or negative are becoming common, but they are only the beginning of what is possible. One of the startups he finds interesting, Recorded Future, specializes in extracting references "next month" or "next week" or a date in the future so that posts can be tagged as relevant to an upcoming event, product release, or deadline, allowing its analysis to target things that haven't happened yet, where there might still be an opportunity to change the outcome.
"Any time you put another dimension into the data, you're improving it quite a bit," Sponder said.
Social media data has tremendous potential as a source of market intelligence, but it's also important to recognize its limitations. For example, one Wharton Customer Analytics Initiative study showed that social media data had a high correlation with offline word of mouth for the automotive industry, but for beauty products there was hardly any relationship.
"The use case may very well be industry dependent," Sponder said. "You may think you are listening to the voice of the customer, but you may just be listening to a couple of irate Twitter followers."
Follow David F. Carr on Twitter @davidfcarr. The BrainYard is @thebyard
The Enterprise Connect conference program covers the full range of platforms, services, and applications that comprise modern communications and collaboration systems. It happens March 26-29 in Orlando, Fla. Find out more.
Read more about:
2012About the Author
You May Also Like
2024 InformationWeek US IT Salary Report
Aug 15, 20242022 State of ITOps and SecOps
Jun 21, 2022