SAM I am: What did you think of those green eggs + ham?

Sentiment Analysis--or opinion mining--typically involves linguistic analysis to discern what people think about something. One's sentiments are usually discerned from the language one uses when referring to some said thing.

Offline, sentiment analysis is something that humans are actually pretty good at. Of course, humans also use visual cues like facial expressions and body language to supplement familiarity and trust when we interpret another's sentiments...and even then we can get it wrong. So it's no wonder that machines are notoriously poor at understanding one's sentiments.

The challenge is daunting: The web and its social media spawn have created a seller's market in opinions (the irony of making this statement on a blog is, well, ironic). And while the difference between fact and opinion may seem clear much of the time, the practical aspects of teasing apart the two in software are not so easily algorithmized (?!).

Enter SAM: The sentiment analysis manager.

Teragram Linguistic Technologies (now a division of the statistical analysis giant SAS) developed the SAM tool to sort through the social media morass and help brand managers understand what its customers really think. It's the latest in a series of tools, such as Jodange's Top of Mind (which we're using in trial mode now), and others by Attensity, Clarabridge, and Lexalytics that attempts to understand who has an opinion, how intensely they hold it, and what trends develop overtime.

So What?

Asking for an opinion is nothing new to marketers...the challenge, as it always has been, is determining what--if anything--you are prepared to do about it. Sentiment analysis software and services are different than traditional forms of opinion seeking in the following several of many ways:

1. Minimize selection bias because they do not require respondents
2. Connect opinions from multiple sources to provide a picture of The Crowd's Wisdom
3. Support diverse visualizations of opinion including heat maps, realtionship graphs and trending charts.

Most importantly, successful development..and deployment...of sentiment analysis tools will enable companies to do what they must to learn from a market that can't be told what to think: listen to it.

Sample heat map of opinion for a client (generated using Top of Mind):

Google Ad Planner

I was accepted as a beta tester when Google's Ad Planner tools were released two weeks ago. You can see what alot of others think about it by checking out their first blush impressions. There's thems that says it replaces the agency and those that sez it's nothing new. Somewhere in between lies the truth.

What does Ad Planner do?

It allows an ad planner to evaluate traffic and audience demographics against specific websites and understand what (Google) ad network options are available for reaching them...online.

For example, let's say I want to reach a traditionally defined demographic online for 25-54 year old females, with bachelor's degree and more than $75,000 income. Google ad planner suggested the following sites (out of a total listing of more than 40 sites that support Google ads and hundreds of other sites that don't):

babycenter.com
perezhilton.com
monster.com
marthastewart.com
healthgrades.com

Apparently, this demographic is interested in babies, celebrities (who may be babies), jobs, healthcare and whatever it is that Martha Stewart does. Who would have thunk that? While these aren't necessarily the insights that make for great campaigns, they do come with some other great Google features. Like the ability to see traffic on these sites.

So here's a quick eval of the good, the bad, and the unknown:

Good:

• I can enter a specific site and see what kind of traffic and the demographic against it.
• For sites that match my criteria, I can see what other sites these people visited and the traffic they support.
• A great way to gain insight into competitive web presenceseses
• The dashboard interface is typically spartan Google and easy to use.

Bad:

• Only sites with sufficient traffic (as defined by Google) will show up...limiting utility for planning against Long Tail sites
• The number of sites able to be researched is far greater than the number who accept Google advertising (which is, perhaps, only bad if you are trying to integrate your planning and buying with Google).

Unknown:

• It is not at all clear where the demographic data comes from...but then Google is likely as trustworthy as any of the traditional demographic data sources (e.g., MRI, Neilsen).

In the end this is a beta. AdWords alone is a $10billion dollar business. As Google enables do-it-yourself media planning--and buying--expect AdPlanner to evolve in connecting marketers with the online world's Long Tail dialogue.

Most of the smart people work elsewhere

Bill Joy (a founder of Sun Microsystems) is attributed to a version of the headline phrase.

Bill Joy

I think he was referring to, as we like to say in the Web 2pointWhatever world, 'crowdsourcing' or in code space as 'open source'.

Though you can certainly see examples of crowdsourcing at work in examples as previously posted ( here ) it turns out that the quote has power for what it implies: that no matter how smart you or your institution are, the good ideas are hard to come by day in and day out...and while once upon a time it was difficult to bring groups of dispersed people together to generate good ideas for solving a problem, the interwebs is here now.

As marketers, we might even find that the always-on, interconnected world enables us to engage (shudder!) our customers in the process of ideas...Product designers, engineers, and market researchers are figuring it out (see Nokia, Dell and even Chrysler). Marketers and advertising agencies--the notorious Big Idea people--might find that the biggest and bestest ideas are out there waiting for them...Fifty-plus years after McCluhan predicted it (though using the wrong tool):

In the 1950's Marshall McLuhan proposed a reality television show in which corporations would present their major problems to a mass audience. "For every expert idea that arises inside an organization," McLuhan advised executives, "the public has a thousand better ideas than you ever heard of."

Tapping into those ideas in the new millenium requires only a willingness to ask and to listen...the means are at hand.

Let's get physical

Back in the day--as those of us who remember hair bands from the 80's like to say--I remember reading about Information Theory as outlined by Claude Shannon. One point that stuck out at the time was the 'Inverse Square' law. Basically this law says that a signal's intensity decreases in proportion to the square of the distance from its source. Extracting the math from this statement one might say, the farther away you are from the sender, the weaker the signal you receive. Ok, that appears to be consistent with the obvious, so what?

It seems that this mathematically provable physical law might also serve as an approximate description of the intensity of a marketing message's impact as it relates to a receiver's distance from the sending source. Come again?

For any message to rise above the background noise that surrounds each of us it needs to have a strength greater than the background noise...and there is a lot of noise in the marketing universe. How can we increase the intensity? We can certainly invest more energy against the broadcast frequency...and reach...of the message in traditional vehicles. Some people even continue turning up the volume on their messages in an effort to rise above the noise.

But perhaps we're looking through the wrong end of the telescope (to mix a metaphor). Maybe we can reduce the distance that a signal has to travel so that its strength isn't depleted by the journey. Mind you, I'm not talking about the physical distance, but the closeness of the relationship between the marketer--the sender--and the receiver. How do you reduce the distance? Here's three simplistic guiding statements:

1. Make the message human...noone has a close relationship with a corporation
2. Make the message local...people will think globally, but they most often act locally
3. Make the message personal...ask, listen, and respond...our differences are where we are most empassioned.

Or we can just try shouting louder.