Isn’t there some sense of irony that the largest analytics company in the world, with the most products and the highest scalable solutions (yeah, go ahead and tell me about better, larger, cuter companies in the comments) had not even entered the Social Media world until now? After all, we are talking about a ready-made solution for a problem looking for one.
The problem is the amount of information organizations have to process thanks to the Social (r)evolution has grown to volumes that we could never had imagined. I talked about this volume when I discussed how to leverage communities with analytics, but something else I learned this week really drove the point home. Not only have we grown to processing petabytes (as in a 1,000 terabytes or 1,000,000 gigabytes) on a daily basis, we have also increased the amount of unstructured data we collect. Back in the days when I was touting Feedback as the do-all and end-all of the world I found out that 90% of the feedback was unstructured. I learned this week that in addition to having grown that number significantly by volume it is now 95% unstructured.
That means that we need to process (in as close as needed to real time) that much data and make sense of it. And that is where SAS shines – massive volumes, quick processing. During an interview I had with Dr. Goodnight, legendary founder and CEO of the company, he discussed how they are moving to do all processing in memory since it has become as inexpensive as storage once was (OK, done name-dropping — although I do admire him). And the speed of processing cannot be matched — heck, it cannot even be compared.
Now that they can process obscene volumes of data very, very fast we are ready to tackle the volume of social noise to extrapolate a signal out of it.
Their Social Media Analytics solution is very complete and CRM Guru Paul Greenberg wrote a detailed review of it that you should read (including my favorite demo-feature, timelines and sliders and emphasizing the company culture – a key aspect to their approach). I cannot think of much they are missing in their attempt to enter the market and become competitive immediately.
First, they have created a unique way to manage influencers. OK, so the fact that they have influencer management is pretty big by itself – acting as a first level filter for the noise captured, but even bigger is the fact that users can manage rules to both identify and analyze key influencers. In other words, the notion of who exerts influence in any specific market or situation is not limited to a pre-established opinion, each organization can determine who their influencers are and analyze data according to the rules. Very significant for a world that is going to depend on reputation and influence going forward.
Second, I was quite taken with the claims for accuracy. No, not because I believe that 90% accurate is significant (actually, any solution can get to 90% with very limited learning and training, it is interesting they can claim it out of the box, but not a significant differentiator). Rather the methodology and tools they have to help improve the accuracy over time. This is a theme that is worth exploring in more detail — which is why I will post more details in about a week when I round up my research.
Third, they have included forecasting and trending tools that work together with their reporting of historical data (up to two years) to determine which way things should or could go, and help plan for it. This is another example of an analytics company taking on a new problem (Socialitycs, as Mike Fauscette describes it) and giving it a total different spin, more inline with what you would expect to do and see in analytics.
I wrote before about how platforms are the new combat zone for enterprise applications, and this release has shown how a platform can play a key role in an organization’s adoption of specific technology. The fact that the power of this solution can be unleashed in a SaaS model, using it as PaaS component, speaks to the power that platforms can bring to the table. We are not just talking about hosted solutions, rather the ability to use them as a platform that leverages all security, data models, business rules and all other information already existing in the organization. That is going to change how we do applications in the very, very short term.
Yes, it is expensive (starts at $5,000 per month after a setup fee of $50,000) but large brands pay more now for market research services that yield less quality analysis. And, yes — it still needs to prove it can do it in massive scale for lots of concurrent customers, but I certainly like what I saw. A lot.Disclaimer: SAS invited me to their event, and paid my expenses. Alas, I am not that cheap of a date and takes more than a few meals and hotel nights for me to “put out” and write what they tell me. But, you already know that since you have been following me for some time. This is all my opinion, and there are certain issues with their solution that need to be worked out — but the overall impression is that they are onto something cool. As usual, time will tell whether I just like drooling for nothing, or got a knack for drooling over the right innovations.