Pamela Meek, Anne Milley, Mark Chaves and Katie Paine
My SAS world and my social media world collided today at SAS Global Forum in Seattle. We launched SAS Social Media Analytics, a product designed for, well, people like me who are trying to monitor and measure the value of social media in medium- and large-enterprises and make decisions based on data.
We had a solid audience of journalists, bloggers, analysts and influencers at the event to watch the presentation and demo. Some of them are people I usually only see at social media events, like Shannon Paul or Alex Williams or Justin Levy or Warren Sukernek or Wayne Sutton. (Despite the fact that Wayne and I live 20 miles apart, I’m more likely to see him in another state than in North Carolina.) The cognitive dissonance was amplified by the fact that I was wearing a necktie, which you can safely assume doesn’t happen at social media conferences.
Following are my notes from the press launch of Social Media Analytics, tidied up a bit. After doing this at BlogWorld and MarketingProfs and South by Southwest, I’m incredibly pleased to have done this at a SAS event.
The panelists were Anne Milley and Mark Chaves of SAS and Katie Paine, the measurement queen, from KDPaine and Partners. Pamela Meek of SAS External Communications introduced the event, and a welcome message from Chris Brogan of New Marketing Labs, who worked with us in getting this product ready.
Chris: It’s really hard to help people understand the value of the efforts they’re making in social media. We don’t understand how to show people that it’s working, so I’m really excited about the launch of the Social Media Analytics tool. It’s a tool I’ve been needing for quite a while, and the companies I work with have been needing. I think the Social Media Analytics tool is really going to change the way we do things in this industry.
Mark Chaves: SAS Global Forum is SAS’ original social network.
Accurate sentiment over time is vital. The Media Portal allows users to understand the impact social activity has on their business. Mark used an example of a bank. The bank may have great financial products but their customer service is lacking. SMA is able to look at the different parts, be precise and accurate to deliver real business value. You can look at a point in time but also over time, and helps you understand how your activities change the results over time.
Mark gave an example of the bank issuing a report that led to an increase in negative activity. The PR team could drill down and see the media source that contributed to the negative comments and where they came from, and the fact that the negative comments came from consumers. We’re talking about taking a look at sentiment not just for your brand, but tied to specific aspects of your business, and analyze them over time.
Anne: Katie, how would you use this?
Katie: The time dimension and the fact that it also addresses traditional media is very powerful. What happens when a negative mention from a media person affects your customers? SMA will let you see what impact it will have over time and how you respond. It’s going to be a hugely valuable tool on preventing the bad stuff but also in seeing how your social media activities work, how the channels work, how they affect Web traffic and conversions.
Anne: Katie, how accurate is this?
Katie: This is better than 90 percent accuracy. I have humans who have been doing this for years and if you can get to 90 percent, you’re really doing something.
Anne: How does the pricing compare?
Mark: We start with an upstart fee to get started, understand the customer’s goals and build a taxonomy of the business to make sure we get accurate results. The second part is an ongoing monthly fee that will vary from $5,000 to $15,000 a month based on the customer need. There are also custom projects that we can do based on the customers’ needs.
Anne: Aren’t we late to market with this?
Mark: Text analytics is the game changer for this industry and we’ve been in text analytics for more than a decade. In partnering with people like KDPaine and New Marketing Labs, we’re able to tap into our global understanding of this area.
Katie: When we started talking about this four years ago, social media wasn’t fully formed. Now major brands are starting to understand the impact. “You’re never wrong, you’re just early.”
Anne: What are the key differentiators?
Katie: The accuracy. People get hired and fired based on the data I give them. This data is accurate and integrated. You can measure real business results, not just activities. And it’s SAS, who understand how businesses use data.
Mark: I’m in R&D so I look from a technical perspective. Customers and their agency partners can participate in making the engine better over time. We don’t just look at sentiment; we look at time and add forecasting and predictive elements. We tend to look at social media as a new data source into data customers already have.
Mark: Sentiment is very diverse across different channels. If you’re a media relations professional, you want to understand the differences across blogs, Twitter, YouTube, etc. Each has different characteristics.
You can also go from a backward view to being able to plan. You can get a sense over time of the volumetric changes, seasonality, etc. If you’re in media relations you want a sense of the noise compared to the useful data, and be able to compare it over time. You’ll be able to see the future and plan accordingly.
You can see, for instance, what discussion forums are projected to grow over time, and figure out which ones you want to participate in.
Mark showed the Sentiment Over Time report. “We’re moving from measurement to management.”
Katie: Everybody gets so upset because an organization like Greenpeace might be attacking them, but now you can see exactly what the impact is, is it changing how people are actually talking about your products in the marketplace and differentiate the traditional media results from the social media and see where to put your efforts.
Mark: We’ve created an influencer score that takes into account key variables: what’s the percentage of their tweets that relate to your topic? How often are they retweeted? We’re able to take these data and apply data mining techniques to help customers understand the actual influence of the people talking about them online.
SAS can also create a social graph view, so you can see where people are talking about these issues.
SMA can create phrase clouds, not just word clouds. The words “short” and “sale” don’t mean much separately, but “short sale” means a lot in a banking context. This tool allows you to look at them as more than just individual words.
Question from Josh Greenbaum: Several times you’ve mentioned the level of accuracy. What is the process by which you can verify that, and can you guarantee a level of accuracy to your customers?
Mark: It’s a really fun process for us, working with customers on this. The hotel we worked with recognized the value of reviews to their reputation. We focused on hotel reviews. After we built some concepts of what specific aspects to measure (the TV, the pool, free breakfast), we develop rules. We manually reviewed the results side by side with the automated results, and got a 90 percent success rate. With blogs it was close to 70 percent, and it can increase over time.
One other aspect of verifying accuracy: There’s the human and manual element, but you can also compare to the structured data on the Web, like the number of stars in a review.
Denis Pombriant: How are you able to filter out noise from a group of people who are purposefully trying to skew your data?
Mark: Noise exists at the initial data capture level and after it’s processed. That onboarding process where we bring on the customer, we vet that data across the customer teams and their agencies. We make sure our filter captures the initial capture of the data appropriately. That will help them learn if there are people who are intentionally trying to skew the results.
Anne: How do you determine the top influencers?
Mark: What percentage are they talking about your topic or brand, how many followers do they have, how often are they retweeted? Add Facebook to that and you have a data miner’s dream. Our customers can take our standard influence algorithms and customize them and incorporate their own data.
Katie: Every client has different measures of success. Unless you can define the results to a client’s measure of success and allow for computer mistakes, automated systems won’t work.
Mark: There is a Web analytics part of the solution that provides the ability to correlate social activity to Web traffic and do some scenario planning. How much social activity do you need to achieve a particular Web traffic goal? This allows for “what if” scenario planning.
What are your customers saying? Can you compare what people are saying in social about you to your customer emails, call center requests and other data from your customers? Does your market perception match up with your customers’ perception?
Wayne Sutton: How well would the product find traffic from geotagged keywords, and in conversations happening in other channels like discussion boards, user groups, less public channels?
Mark: The approach we take is to begin with the end in mind. We first start with the strategy. Because we know so much about Facebook data, we can take advantage of the richness of the data. We’re also partnering with Effyis. There are new media sources that allow us to be richer.
Katie: When it comes to geotargeting, with human testing and variation the answer is yes, but it’s a lot harder than you think. It’s not whether you can figure out where they are, it’s how you can get rid of all the other junk that identifies the place. It’s a detailed process that uses the human element to refine the results.
Anne: How are measurements in one department affecting others, like marketing information impacting PR and vice versa?
Mark: Conceptually, integrated marketing is harder than it sounds. Marketing is siloed: marketing, PR, etc. We look at it from an integrated marketing perspective. If you start with the end in mind (drive traffic to our page) what are the components and how do they work together? You can’t do planning unless you understand the ebbs and flows of traditional media.
Alex Williams: Is this on-premise or in the cloud?
Mark: It’s an On Demand solution that we can adjust as the customer needs changed. It’s a solution that’s delivered as a service.
Shannon Paul: When you monitor you run into the same people all the time talking about your brand. What’s the potential for integrating anecdotal analysis into the larger data points?
Mark: Our customers share with us the data they already have, so their CRM data can be integrated.
Katie: Anybody who already has a system set up can add their data into this.
Mark: SAS has adapters to Salesforce. Com and other CRM teams and has been years. What if you have a group of moms who wield influence around the brand? You can tie that to the CRM system so you know their customer data as well as what they’re saying about you. You can slice and dice so that you can treat customers differently based on their interaction with you and their value as a customer.