Last night I attended a Social Media Analytics Networking Reception at the SAS Global HQ in Cary, NC (I have been a guest of my American colleagues for a few days whilst I waited to find a flight home). It was a great evening with lively and amusing conversation.
After a short and highly informative demonstration of the recently launched SAS® Social Media Analytics, there was a question and answer session and one particular question prompted an off-the-cuff response from me because I had heard something very like it only a few days before.
After being told that one of the key differentiators of Social Media Analytics was its ability to forecast future changes in sentiment, the question was (and I paraphrase slightly), "How can any system predict how people will feel and what they will say about an event that hasn’t happened yet and may not happen on a schedule?" Good question.
My response was based on separating the cause from the effect. I said, "No system can predict when a random stone (event) will get thrown into your pond; however, once the stone has entered the water, you can predict the behavior of the ripples (consequences)." With the kind of advanced analytics that SAS uses, it is possible to build models that will predict how the ripples of sentiment will spread once the originating event has taken place. Continuing the analogy, you can even take into account how big the stone was (how severe the causal event) and whether it hit a backwater or right in the centre of your pond.
To stretch the analogy even further, would you like to know almost immediately whether an event is going to cause a tsunami or a ripple? Or would you rather wait at the bank until the surge reaches you (and by then has reached everyone else as it radiates out from the center)?
When it is my reputation on the line, I would rather know whilst there is still time to take some kind of remedial action, than after I have been inundated by the tide of consumer sentiment.
5 Comments
Peter - I enjoyed meeting you finally, and sitting next to you at our local social media gathering. I liked your analogy. There's another point I'd add: Beyond the ability to predict how the ripples will spread when an event occurs, the predictive element in the SAS application can help you notice that an event has occurred or is occuring, before it might otherwise become obvious. Maybe more like a geyser than ripples, it would let you recognize temblors or other early warning signals well before the water bursts through the surface, and either cap or divert the flow before it causes trouble. It would do this by "noticing" that patterns were NOT occuring as predicted.
Great event the other night, great comments here, too on this post. My colleagues and I were discussing this exchange, also.
The problem I think I'm having with the example that was offered that night - Toyota's recent brake issues - and the world "Predictor" is that no tool can predict a positive or negative social reaction to an event that has not yet happened. In hindsight, I don't think the Toyota example was applicable to what SAS's SMA solution does - SMA certainly CAN predict seasonal responses in sentiment: I sell backpacks for school - clearly sales will go up in August for Back To School time of year interest. But I don't think SMA can "predict" (to your original point Peter) that two punks at Domino's are doing horrible things to a pizza, filming their antics and then uploading the proof to YouTube. Instead, SMA may show, as you describe - the immediate effects of the impact as the stone is cast into the water and the potential effects of the ripples as the media spreads through that brand's influencers (to Diane's point.)
I hope I have that right?
What I wanted to better understand - is how SMA can integrate with, or impact website analytics and play nicely with other marketing data sets. Hopefully, our agency will get to play with it a bit more soon!
Ellie clearly gets it, and her analogies were spot on. We of course can't predict the next Domino's fiasco, but we can do an effective job of projecting the impact of that fiasco as the social chatter around that begins to pick up steam, using both the immediate feedback, and the conversation history baseline to inform those projections.
On website analytics, we recongnize that the two go hand in hand, and have integrated our web analytics software (and associated data collection capabilities) as another module that allows you to view website impact along side the Social Media Analytics insights. We can also accept data sourced from web analytics providers, which can be funnelled into whatever SAS software you want to bring to bear against that data.
As for other marketing systems, we also have hooks into CRM systems like salesforce.com, which may have customer transcripts or other sales and marketing data relevant as an overlay to give a much richer view of this space.
This specific Q&A highlights what we hear again and again, that companies (and especially agencies) are looking for more than just reporting, and what you've seen from SAS in the launch of this solution is a reflection of that customer feedback. Thanks for your comments Ellie.
Regards,
John Bastone
Global Product Marketing Manager
SAS Customer Intelligence
Hello Diane,
Nice to meet you too and I enjoy the social evening.
I agree that you can use Social Media Analytics to get an early warning that there is a building change in sentiment. Interestingly, it's not always based on a 'seismic' event :: just one case of poor customer care could be the instigator behind a negative campaign - one that you might not be aware of until it has picked up a head of steam and heading your way at full speed.
There used to be an (old) adage that one dissatisfied customer would tell 10 other people; well now it can be 100, 1000 or even more (depending upon the 'weight' of the complainer. Also, those people can tell another 10, 100 or 1000 each - before you know it, your customers can be swapping 'horror' stories about your everyday business practices - things you never knew were so critical to your reputation.
In my mind, Social Media is the game changer - Social Media Analytics is simply the first powerful tool that helps us understand and play by the new rules.
Peter,
Great to meet you at the event last week. Thanks for clarifying the role of predictive analytics in SAS Social Media Analytics tool. I think the analysis of past behavior allows us to plan better and anticipate predicted fluctuations. We will always have to deal with unexpected variances to our social web presence as it's impossible to predict customer behavior 100% of the time. However it was nice to see a strong analytics tool that does a better job of packaging the large amounts of data and narrowing down trends to specific phrases versus piles of keywords.