Friday, April 08, 2011

A challenge for the way sentiment is measured

It occurred to me when watching my twitter stream fly by (I follow close to 3000 people) that I rarely see anyone share that they have been at an 'unawesome' event - or had the bad fortune to spend time with really dull people.

When we say we've had a good experience we are showing off a bit - "look at me - aren't I good at making the right decisions". We don't like to be seen to be making the wrong decisions. It's part of us retro-fitting a model of rationality on to our more likely Herd behaviour.

Which I think is one of the reasons we tend to be more positive about the choices we make compared to those imposed upon us. No need to post-rationalise when the decision wasn't ours. (image courtesy niznoz)

Of course, my friend Mark Earls would argue most decisions aren't ours, they are more driven by physical context and those around us than we'd like to admit. But the kind of 'not my choice' I'm thinking of is more in line with being negative about the train service you have little choice about taking, the utility or bank etc where the cost of switching appears high (in effort terms at least).

Which is all a long way of saying we big-up what we choose for ourselves and more easily criticise what we don't. There's a lesson in there for building positive sentiment about brands and products and change for the better. And also a challenge for the way sentiment is measured and valued.

Sentiment isn't a standalone value or a necessary consequence of product, positioning or recommendation alone. Good product does not (always) = good sentiment. Good product + 'I chose it myself' may ampifly the positive. Good product - 'someone chose it for me' may dampen it.

For me this is another example of the questionable value of sentiment analysis in and of itself. It cannot be a stand alone value, it has to be measured in the context of a number of complex factors - the purchase cycle, the relationship of 'peers' in the distribution of recommendation. And then its measurement must be put in the context of what a shift in sentiment here or there results in elsewhere. Folk being a bit happier about your product means nothing on its own. That playing out in rising sales, more engagement in co-creative processes (both active and passive) etc (in other words things that deliver real value) is where the bang for the buck resides.

And I haven't seen a tool yet which does even half of that.

The world is full of data. Always has been. We get access to more of it than ever before - thanks to digitisation of thought, emotion and opinion through the likes of blogs, forums, texts, emails, twitter and facebook.

Making useful, innovative efficiency driving sense of it is the key. And that remains a very human art.
Enhanced by Zemanta

1 comment:

  1. I've been thinking about this recently and just talked and wrote about it here: http://www.nixonmcinnes.co.uk/2011/03/31/poking-insights-why-do-we-lie/

    Interpretation of sentiment analysis still needs context: who is saying something (most people don't), where and when are they saying it (in bed, in a shop, watching TV), what is the purpose (for me, for her, for them).

    It's trite but it's the contextual understanding that leads to insight and the insights help create the value.

    ReplyDelete

FasterFuture.blogspot.com

The rate of change is so rapid it's difficult for one person to keep up to speed. Let's pool our thoughts, share our reactions and, who knows, even reach some shared conclusions worth arriving at?