The 2014 World Cup was billed by many as the first truly social world cup, with Brazil’s semi-final humiliation generating 35.6 million tweets and becoming the most tweeted event in history.
Of course in reality, World Cups have always been social, drawing in people who don’t even like football to watch and talk about it with their friends. What the "first social" billing is about is the way this is now happening, and the mass of conversational data created for marketers to use in planning and implementing campaigns. This is really the "first data World Cup".
The scale of the World Cup makes it an ideal test bed for understanding how this morass of conversation data can be used in real time marketing.
To do this we set up what we call a Loop Room – in essence a place where we can pull key data feeds together in one place, to inform decision making. We tracked social conversation, search trends, video sharing, and game stats across the duration of the tournament.
One of the key challenges to real time marketing is the speed with which decisions need to be made. For brands that wish to capture the zeitgeist, real time marketing necessitates real time decision making which can act as a limiting factor on creativity.
But if you knew what conversations would exist around a topic in a couple of days’ time, you would have much more room to manoeuvre and plan your activity accordingly. "Pre-reactive" planning if you like.
We partnered with BlabPredicts, a predictive analytics company, to try to do just this.
So, what did we learn?
Predictable things are easier to predict. This may sound like an extreme statement of the obvious. But predictive conversation analytics is dependent on a level of conversation already existing to create a forecast.
Conversations that are steadily building before a known event tend to be easier to predict than conversations about an event that has only just occurred.
Blab had most success with conversations leading up to a specific event, for example discussion of the state of the Manaus pitch before England’s first game with Italy.
'Events, dear boy, events'
Or so Harold Macmillan is alleged to have said when asked what is most likely to blow governments off course.
Predictive conversation analytics suffers from much the same problem. It's not possible to predict conversation around specific unknown events, such as the Suarez bite, before they actually occur.
The role of a tool like Blab is to frame the likely longevity of conversation around events when they do occur, rather than predict them in the first place.
How long will conversations last?
One of the challenges in using real time conversation data to shape campaigns is knowing how long a topic is likely to trend and whether it is worth a brand latching on to.
Blab indicated correctly that the Balotelli sticker album story which briefly trended was unlikely to grow.
Decision making support
If predictive conversation analytics works best for the predictable, what is the point? Our experience showed that its role is to give guidance and support or challenge hypotheses that you already have.
It allows you to compare the likely relative growth in multiple potential conversations simultaneously.
So, you may have expected conversation around Rooney to grow ahead of England’s first game but it might not have been clear whether Sturridge or Sterling would be capturing the greatest share of conversation.
This is a field in its infancy. To predict what the world is likely to be talking about around the World Cup sounds fantastical, but the accessibility of vast volumes of conversation data and improving analytical techniques means that it is increasingly possible, albeit with limitations.
As more brands look to latch on to the zeitgeist and adapt campaigns in real time, the ability to predict what people will be talking about, even 24 hours ahead, will become increasingly valuable.
By the next World Cup, if not the Olympics before that, forward facing marketers and their agency partners will have strategies or scenarios planned out so as to be able to deploy and adapt at the speed required to keep up with the most fluid of conversations.
Jeremy Pounder is client director for business planning at Mindshare