Social media mentions are the tip of the iceberg, they are fun and can offer a high level overview (i.e. http://apps.facebook.com/usa_gnh/) but they dont consider variables which could affect the social media mention and should therefore be considered part of the holistic process of measuring sentiment.
Sentiment is a measure of consumer confidence, a form of Mass Opinion Measurement.
Traditionaly it was measured through telephone interviews on a regular (i.e. monthly) basis giving a ‘near-time’ measure of consumer attitude towards something broad, like the economy.
Traditionaly the inputs were the answers to a series of questions asked to a large number of participants over the phone, such as:
- Are you happy?
- perception of the economy over the long term / short term?
- perception of you current financial situation?
- perception of your current job security?
Get the metric right (that’s a whole other thing) and the outcome is near-time sentiment.
In just the last twelve or so months of radical change in the social media landscape, a new way of measuring sentiment has been born. We can tap into the openness and the API’s of twitter (thats the major change in the last twelve months), blogs, user generated comments, photo and video sharing sites etc.
Without asking people, we can see what people are saying and collect data on where they are geographically located, how old they are, what sex they are, what they do for a living and other relevant information
We can take this information and filter down into granular detail, cross reference against current data on the local economic situation; unemployment, average wage, crime etc. We can take into account the local ecology too – like the weather, the landscape (i.e. hills and stuff – relevance dependant on the product / service you are measuring)
We could also look at historic economic factors too, like quality of life index now and in the past, cost of living etc..
We can then build an algorithm to collate all this data and the outcome is realtime sentiment, sympathetic to external variables (i.e. the current season ifs its soup, the amount of hills if its a bike, the waves if its surfing etc.) If an FMCG can get this real time visibility on sentiment they can react quickly to leverage the good things and weed out the bad things.
Issues:
Reliability – as always, the simpler you make it the more effective??- In its simplest form, group twitter comments together for an indication of how we collectively feel about something.
People who do it:
http://www.sysomos.com/reports/topbrands/
nielsen buzzmetrics
.. I dont think any ‘out of the box’ solutions consider all the data I mentioned, but its the future, I promise.
How do we make money from it:
Not in providing the service but in facilitating this drive towards more refined analysis and scooping the additional spend on interactive. None of this kind of monitoring is available with trad above the line advertising. Stuff like easily accessible sentiment monitoring will help drive the trend in channelling more money to digital.
Tom Burford
Business Manager
