One thing people do on the web is create a huge and only slightly disorganized knowledge collaboration space. On a daily basis, just in the English language alone, they create about a half million blog posts telling us all about themselves, their beliefs, attitudes and opinions. What’s not to love about that! So how to make sense of it?
Most marketers try to understand this collaboration space through sentiment analysis. This gives you a snapshot – are people positive or negative about my brand, product or company. I recently wrote this about sentiment analysis on my own blog:
“The problem of sentiment analysis has generally been something like this – people create double negatives such as not unkind, to mean a positive. And then there is the general sense that people are ambiguous in how they express what they mean.
My problem with sentiment analysis is neither of these. It’s the fact that companies need to connect with people in different ways rather than through a scaled up focus group or opinion survey; and that people are extremely granular in their sentiments. Companies need to know people’s ideas and beliefs at the granular level not at a polarized level.”
Clearly I erred here because sentiment analysis does not yield as much information as a focus group or opinion survey – in both those cases there is granularity, though in my view the results are too easily manipulated or geared by the interviewer.
For example – and anyone in marketing can answer this: how often are brand recognition surveys consciously or unconsciously biased because the company paying for that work needs to show improvement?
What the web offers is a way to get at volunteered opinion and attitudes.
To that end I and a couple of colleagues have spent the past year developing the global attitudes project, which I hope to share with members of the nGenera/Wikinomics community over the coming weeks.
To do good web analysis means we really have to accept the validity of what people are saying, and accept the web as a bona fide collaboration space where people collectively give meaning to life in the 21st century - an unparalleled exercise in meaning creation.
In the past meaning creation has been the job of intermediaries – philosophers, trades unions, politicians, writers, teachers, academics, and of course market researchers. But now we’re doing that for ourselves.
Look at it this way – how many studies do you know where researchers have interpolated meaning for us? Most trend studies and projections do this – telling us we are becoming more individualistic or more short term in our actions, or more multitasking, etc. Actually on the web people are telling us what they are and what they believe.
I think the narrative accompanying this new description of what people are about is something like: we still believe in human progress; we don’t necessarily believe society as it is now organized is a vehicle for that progress; we believe in individual betterment but we also believe we are part of an organic “world ecology” with life cycles that rise and fall. We might be in the fall phase so we need to work harder at renewing our human ecology.
I can’t present all the evidence for that in a short blog post but what I want to do is give a small evidence base for one term and one term only.
Before I do that – why is it important? The answer is the Web provides us with deep insights into how people are thinking. You have to care about that if you are in business. You can short circuit that understanding by using sentiment analysis – but be aware it is a short cut and lacks essential insights.
So: Throughout the recession people online have been making increasing use of the term betterment. We know that through a variety of data collections we’ve executed.
We analyzed the use of that term semantically, i.e. asking what meaning are people attaching to it. Here’s a graph:
The graph actually shows you the terms people attach to the term “life” when you use an evidence base made up of blog posts referring to the term betterment. Why does it show “life” at the core?
Because that is the term most used in a selection of blog posts that top the Google blog search returns for the term betterment (as of Mid May 2010).
Sorry if that is confusing. This is an early result and we may yet refine the technique, data, analysis and visualization. I wanted to share it as an example of what can be learned from the web when you go beyond sentiment and polarity.
What it means is that when they discuss betterment, many people in this sample use the term “life” at least once. They use “life” more than they use any other term (there is a refinement to rule out the possibility that people always shoot the breeze about “life” and that rather than betterment is a driver).
And when you look at the semantic family of terms around “life” you see the graph above. It’s mainly constructed of terms that evoke relationships.
What is not on here because the terms are not present enough are terms like God, i.e. people attribute more to the core human temporal aspects of life than they do to religion.
We began this whole area of study with a review of about 20,000 comments on “recession and recovery”. In that study we discovered that people in blog comments tended to interpret recovery and recession in terms that evoked the human side – i.e. by referring to people. In parallel we collected data on how the financial press reported recovery and recession and found, by and large, they focused on institutions.
I make that extra point because the two examples taken together probably tell us we need to retune how we speak about marketing, the web, social media and analytics. Once we look at the world from the vantage point of this constantly emerging philosophy we have to put people and realtionships first. Marketers need to engage with this non-polarized way people express themselves. More at a later date!