Are we Good or Bad? Happy or Sad?

Here I am back at work extolling the virtues of a 'stay-cation' by having a Britain-acquired suntan without the need for flying (hah!), and showing my age with a quote from a great Tina Turner song.

I always enjoying working with the Marketing department, and most of all I enjoy chatting with their Insight team - a group of people who find out the customer perspectives on our business.

They have shared with me some work they have been doing with social networking data to see if they can use this 'crowd-source' of information to find out what people think of Tesco. They are even looking at how our customer service staff might be able to interact with customers who are having problems with us.

The team having been working with a company called Right Now who specialise in monitoring clouds of social media to uncover what good and bad things are being said about us. One of the most interesting reports we get from Right Now is a distillation of Twitter feeds where customers mention Tesco. Importantly, each 'tweet' is analysed to discover the sentiment of the message and is marked on a grade from positive to negative.

Seeing as how Twitter feeds are entirely open and in the public domain, and tweets are available in their untold millions, it's a perfectly legitimate source. However you may argue that people of the sort who use Twitter and splurge their thoughts on that network may not exactly be representative of the Tesco customer base. You would be partially right - after all, Twitter tends to be used by a younger demographic with enough wealth to run a mobile smart phone and/or have easy internet access wherever they are. However the sheer weight of numbers allows us to get an overall sentiment and, more importantly, achieve early warning of any change in sentiment (more negative or positive) for some reason.

For (contrived) example, supposing we ran out some important product that people expect us to sell, for example bananas. Tweets suddenly appearing with negative sentiment that bananas are missing could, in theory, allow us to take action earlier than discovering it later. If tweets reveal that a group of stores are affected, we could home in on the problem more quickly. For example, it could be a broken-down lorry that planned to visit these stores and its stock needs to be transferred with greater urgency than just using 'normal procedures'.

Bigger business challenges can also be uncovered - take these recent negative tweets about Tesco Clubcard that uncover miserable cashiers, problems registering a clubcard or changing address, and an upset 17 year old too young to own a card (click image to read text):

On the other hand, there are plenty of positive tweets - this excerpt uncovers the joys that can only come from Clubcard Points and Clubcard Vouchers when used with various deals. This shows us that we really scored with these offers and its' good to get the feedback. It also reveals (again) the importance of reaching 18 years old so as to be allowed to have a Clubcard. (click image to read text):

A real quick win when it comes to finding out what key words (and by inference, sentiment) people are using when they Tweet about a subject is to use Tweet Cloud. This service finds out the top 30 or so words being used alongside a search word in tweets. The more commonly a word is used, the bigger it is in the cloud.

Here are two clouds taken this morning (Mon 19 April 2010) that show an intriguing difference about people who use the word Tesco in their tweets, and people who want to let you know they are talking about Tesco by using the hash-tag #tesco.

First, a 'Tesco' cloud (click here for a live cloud):

So I'm seeing stuff on the Iceland volcano, a couple of competitors, voucher (no doubt Clubcard vouchers), words such as 'clothing' and 'entertainment' - both are being promoted at the moment so its nice to see people reacting - and only one expletive.

Compare this to the cloud of tweets where the authors want you to know they are talking about Tesco by using '#tesco' (click here for a live cloud)

Here the sentiment is more business-like - about jobs, competitors, and the occasional politics.

The insight team are adopting the strategy of listening at the moment, and as you can see there is much to listen to on the internet's public domain. I'll keep in touch with them as they plan their move into the 'interaction' stage.

In the mean time, check your brand sentiment. Since you and your brand have been together,

as Tina sang, times are good or bad, happy or sad?