Posts Tagged ‘SAS’


SAS introduced its Social Media analytics product today and, given that SAS is SAS, it will serve to give some new legitimacy to the field.  Social Media analytics has been around for a long time, probably as long as the Web minus a day or two.  But that doesn’t mean that SAS can’t add to the conversation in some important dimensions.

For instance, two of the major constraints on analytics have always been CPU throughput and memory.  The more CPU cycles you can throw at the problem of finding a needle in a proverbial haystack the faster you can find that needle.  But CPU only works on the data in memory and if you need more data from disk each seek slows you down.  So SAS has done a lot to provide the highest performance possible by throwing CPU and memory at the problem.

Last night at the opening keynote, Dr. Jim Goodnight, showed what this can mean.  He took an analysis process from a bank that had taken eighteen hours to load and reduced it to a couple of minutes by putting the whole shebang in memory and dividing the task among just over a thousand processors with multiple cores.  The importance here should be obvious, that massive amounts of data that a large enterprise might typically generate or have generated about it can now be analyzed in time to make the result relevant for ongoing business processes.

The simple performance demonstration preceded today’s announcement of SAS’s social media analytics for a good reason.  If you add together high performance and the very large datasets you arrive at a useful solution for corporate marketers trying to make sense of the twitter-facebook-blog-and-other-social-media data streams that are a fact of life today.  As the demo at the press conference made clear, it’s nice to know that sentiment may be up or down but it’s even better if you can analyze it by source, season and other parameters.  The result may enable a marketer to pinpoint a particular article or post and determine the most effective course of action.

In the right hands, Social Media analytics can also help you understand the unintentional information that’s given off by your competitors whenever they enter the social web or when the market reacts to something they do.

SAS is offering its analytics as an on-demand package but it’s not simple or something that you buy this morning and start using this afternoon.  The company has a well thought out process for on-boarding customers and sticking with them over time to mentor, coach and occasionally perform consulting projects.  This all seems very reasonable.

Several smaller Web marketing and marketing analytics companies have taken the same approach lately of providing product and ongoing service and it’s a business model that I think will become more common over time.  We’ve spent decades trying to make analytics simpler to use and the results have been good.  But the reality is that it will always be something close to rocket science to understand and use tools that predict what people might do with a given amount of information.  Also, the marketplace is changing.  In a zero-sum marketplace, like the one we’ll be in for a while, it is shrewd for a company to cross sell service rather than another product.  Selling service further cements the bond and enables further discovery leading to more cross and upselling.

So the net of the SAS announcement.  Interesting that SAS has put a stake in the ground in social media analytics, their enterprise customers will appreciate it and I think it’s important for the continued growth of the company to offer a service like this.


One of Salesforce.com’s challenges in driving Chatter’s acceptance comes from positioning it for the buying public.  That’s a tall order since the company is simultaneously trying to establish a new product and its category.

The product and the category are classified as social networking and leverage the wisdom of crowds — James Surowiecki’s idea.  But Chatter is unlike any of the products that may help a group come up with the answer to a quantitative question of the type, How many jellybeans are in this jar? which Surowiecki uses a lot in his 2004 best seller.  Nonetheless, Surowiecki does address the Chatter problem as one of coordination and the wisdom of crowds is a good approach to dealing with coordination with some caveats.

To review briefly, the wisdom of crowds is almost self explanatory — the crowd is smarter than any one of its members and crowd wisdom can lead you to some astonishing revelations such as the number of jellybeans in the mythical jar, the best route to work, which styles will be popular in the fall and which won’t.

The Chatter problem is different from all those examples because the group being sampled is internal to the organization and the customer problem the organization is trying to address may be unique.  Regardless, there are three attributes of a successful wisdom of crowds strategy that all approaches seem to need according to Surowiecki — diversity, independence and decentralization.

The three strategies work remarkably well for the jellybean problem and they fit the Chatter problem well too.  Briefly, diversity means getting input from as many sources as possible, not just the smartest people in the room, but everyone.  Smart people tend to think alike and a creative spark can come from anywhere so the more the merrier.  Also, diversity means capturing bad ideas as well as good and the wisdom of crowds makes it possible to let bad ideas cancel each other out leaving you with the good stuff.

Independence means letting each actor in the crowd do what he or she does best without attempting to influence them unduly.  Many of us think nothing of begging friends to vote for our ideas in a social forum as often as possible, but as soon as we do, it wrecks the idea of independence and therefore the whole wisdom thing.  Finally, decentralization is tied with independence in this example, your begging produces a command and control hierarchy which is responsible for the wrecking.

So, what about Chatter?  Chatter neatly implements the three primary social strategies of crowd wisdom for the benefit of the organization.  First, there’s diversity — everyone does a job, their job, not someone else’s.  That means support does support and sales does sales and when there is a customer problem involving a deal being held up by a support issue, everyone does their part.

Chatter helps this process by opening up lines of communication — and ad hoc coordination — so that sales knows what support is doing and vice versa.  Importantly, anyone else who wants to know and lend a hand can also have a ringside seat.  This is critical because other independent actors, perhaps a manager, VP or C-level executive, can shine a light on the situation too.

Finally, the critical piece is decentralization.  In a Chatter crowd, no one has to be a dictator, taking control of a situation and directing people to do certain things.  People do their jobs and that turns out to be enough.  The secret ingredient to this decentralized approach is corporate culture.  A culture that says, do what makes your boss happy might have a tough time benefitting from Chatter.  But a culture that operates on a prime directive approach — like Google’s or Star Trek’s — will do well.

The prime directive can be as simple as “Don’t be evil,” or “First do no harm,” or more practically, “Be good to customers, and always to the right thing,” is all that’s needed.  An employee operating along the lines of the culture’s prime directive should be praised and rewarded regardless of outcome, which will usually be fine.

I am at the 35th meeting of the SAS User Group this week in Seattle.  SAS is one of those gems of a company we all wish we worked for.  In fact, they just won an award for being the best company in America to work for, an award that goes with similar awards from all over the world.  I haven’t spent a lot of time following SAS since I worked at a competitor more than a decade ago but what I am learning here is truly remarkable.

I don’t know what SAS’s prime directive is but I know that every employee does.  You can see it in their eyes and hear it in their conversations and it is definitely pro customer.  What makes SAS such a gem is that it is privately held and unlike many private companies this one has revenues north of $2.3 billion and it reinvests more than 20 percent in its products and services.  And it appears to me that founder and CEO, Dr. Goodnight, has a prime directive that encompasses customers and employees that I assume goes like this: treat employees well and they’ll treat customers well.  It’s a strategy that has worked well for more than 34 years.

SAS grew into a powerhouse without social media — though today it is introducing several products in that realm.  Most big companies are not private and they have shareholders to keep happy as well as customers and employees.  I can’t help but wonder though if applying crowd wisdom through a product like Chatter might help many other companies to apply their prime directives in ways that help them keep customers, employees and shareholders happy by just doing good.  The prime directive idea is tied to corporate culture that that may be the best indicator of whether or not a company is Chatterizeable.  (Is that a word?)