White Noise
About four years ago, CoreData and the good people at Google were working on the idea that the behaviour, comments and ideas circulating the internet were a pretty interesting proxy of what was happening in the world and how people were feeling.
The theory was pretty clear – if we could produce a way to grind up and analyse all the internet chatter, on things like Facebook, My Space, Bebo and Twitter, then we would have a great proxy for what people are thinking.
Google liked this idea and we played with it for a bit, we built skimmers and scanners and started to suck in data about what people were saying.
But it turns out, what people are saying isn’t that interesting, what matters is what they are doing.
Let me give you an example; during the time that Qantas was shut, the Twitter comments went ballistic (see below).
According to the feed – Qantas was a vile company that you would never fly with again and the brand was irreparably damaged.
The purist version of what a company is worth and what it’s being affected by is its share price, which at the time of the PR crisis climbed steadily as you can see below.
The people who were making decisions about what the grounding meant to the business were unfazed.
This was the strength of Google’s argument – what people talk about on Twitter and Facebook isn’t real – it’s just noise – and that trying to divide intent, direction and information from noise is effectively not time well spent.
They claim – and we happen to think they are right – that what people say on Twitter isn’t that important, it’s what they do which is more interesting.
With that in mind, if you want to see what the buying intentions of a country are, you can simply use Google trend data to see what they are searching for.
Below is the searching volume in Australia for Rolex. It’s an interesting chart because it shows that each year in November and December, a time of great gift giving, the interest in Rolex spikes.
During the pre-GFC years there was strong demand at Christmas and even the little spikes for the American and Australian bonus windows, but all that now seems to have evaporated.
If you were someone who predicts the future of a stock price by looking at the chart trends – this would tend to indicate that the future for Rolex this Christmas isn’t bright.
Rolex Search Volume Chart
It turns out that what happens on Twitter isn’t half as useful as understanding what people are actually doing.
3 Comments on “White Noise”
Post a comment
Ryan Tracey says:
Good data. Another reason for the difference between what is said and what is done might simply be that it involves different groups of people. For example, while plenty of people eagerly dump on Qantas via Twitter, clearly plenty of other people are unfazed or even supportive of the brand and so continue to fly with them.
Julie Spencer says:
You dicusss that what people are saying isn’t that interesting it’s what people are doing that’s interesting, true, but I wonder if google search volume is the greatest proxy for what people are actually doing. For example, I frequently check Hermes watches & tiffany’s online, but the likelihood of me purchasing a 3k watch or a 12k necklace in the near future is pretty slim. What would be really interesting to see is the search volume index tracked against the actual sales volume for that product to see if volume is related to what people are actually purchasing ..
Social media skeptic says:
Timely article given Qantas’ latest media storm with the whole ‘Qantas Luxury’ hashtag promo backfiring. Just goes to show the perils of social media and how quickly things can snowball.
The video parody below says it all:
http://www.youtube.com/watch?v=QTCwPlWzZnQ