Why Wordle drives me nuts

In the past three years use of Wordle has spread like wildfire, profession by profession. Right now I have evidence that it is taking the HR profession by storm and late last year was at a conference of school guidance counselors where Wordle charts on PPT create a ripple of oohs and aahs from across the audience. How did the presenter create those?

In some respects I love Wordle, and I love the fact that IBM gave its developer free license to just go ahead and develop the thing, and the way they really put it in the public domain.

But in market research, a word cloud is a poor substitute for real analysis. At best it does a word count so you can see what words got used most. I guess that’s fine if the question is something like: “What is your favourite brand of coffee?” From the massive choice available, the open-ender might produce a cloud where some brands emerge as more dominant than others. But why not just do a bar chart based on mentions?

No, what drives me nuts about Wordle is the way too many researchers somehow feel that by generating a word cloud – which is akin to throwing a pack of cards on the floor – they have somehow “analysed” the content.

Analysis?  To illustrate how good this is as an analytic tool, I’ve created a Wordle to help us analyse some numbers. Hey, it works for words, so this should be brilliant.

WORDLE NUMBERS 2
Hey. This is a good analysis!

What does this tell us?  What numbers aren’t mentioned?  Are there themes at work here? Who can tell? All the process has done is throw them on the floor in a game my sister used to call “52-Pick-up.”  Well, this is all it does with words.

Analysis ought to involve more thought than this, and in a discussion I started in the LinkedIn group Innovation Insight, one member, Shannon Gray from Nashville, discussed how she develops word clouds to go just a few steps further by first of all using Excel to help look for common links between words “unfair – price” for example – and then running word clouds. She recommends Tagxedo (which is very similar to Wordle) because it offers more control.

I do think text analysis goes way, way beyond Wordle’s scope, and the pity is that too many people have been caught up in the novelty of word clouds (broadcasters were big on it last year, but it has fallen from favour with their graphics teams) without thinking: wait – is what we’re doing really adding value to the data?  If this is all we’re doing to verbatims, then it might be better to simply hand over the sheaf of responses to the client and say, “Hey, here’s what your customers are saying.”

If you wish to check out Tagxedo – CLICK HERE

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s