Big Data At The Super Bowl

How does the big game use Big Data?


When people think about how data is used in the Super Bowl, it is generally around the prediction of the winner.

In the weeks leading up to the game, hundreds of companies come out saying that they have the data to predict who will win from the information that has been gathered throughout the year on each team.

With the amount of data available across the teams this is relatively simple, it is possible to see which is the ‘better’ team across a number of metrics, but what it cannot do is give absolute accuracy, hedging bets one way or the other. Many of these companies also make these predictions not for any real benefit except to gain increased exposure.

Many of these models are attempting to simplify what is essentially a complicated sport. This is not just in the formations, injuries and plays, but a number of separate factors can have significant impacts on the way that people or entire teams perform.

Big Data allows people to not only see the number of successful passes, runs or interceptions, but to compare these with other aspects that could have a significant impact on the result. This could be anything from the impact of elevation, when the game is being played or the weather conditions.

The problem with these predictions that come out days or weeks prior to the game, is that there is no way of knowing some of the most important variables, meaning that Big Data may help to predict beforehand or even at the first whistle, but with variations changing rapidly throughout a match, it is almost impossible to get an accurate result.

In fact many companies boast about getting 60% accuracy rates in their predictions for pre-season games. A 60% chance means that although they may get most right, they are still getting 4 in every 10 wrong, which is not a good prediction rate.

The real data driven aspect of the Super Bowl, comes from the advertisers.

This data is also not just about what people are watching on their screens during the game, but what they are doing in the weeks and months leading up to it.

Web analytics and the data gathered from people’s browsing history and purchases, will help advertisers to identify who is interested in what and what is likely to have the biggest impact on them during or before the big game.

This could be from seeing that people have been spending more time than average searching for places to watch the game or looking at potential ways to entertain people in their homes. These are the people who are important to the advertisers, the ones who are going to not only be most likely to pay attention to the ads on the TV, but also to spend prior or during the match.

This information in the hands of advertisers is far more valuable than a flashy ad for the wrong product.

If they can identify who is going to be buying what and when in preparation for the Super Bowl, then they have the ability to offer them what they want to see at the opportune time to maximize the chances of selling.

It can also be incredibly useful in planning what they are going to be advertising in the big budget ads that run during the game itself. If they know what is going to be popular amongst the people watching the game, then they can advertise the best possible product in the best possible way.

However, the kind of data generally looked at from both game prediction and advertisement targeting, tends to be historical.

Increasingly companies are looking at real time analytics to either measure the success of advertising in other formats or to react to live events to target particular segments.

This is generally done through social media analytics, which allow companies to see what people are thinking about their products and adverts and also to see what is being discussed at any one point in the game. Therefore, if something happens that is relevant to a product or service that a company provides, they can utilize the real-time analytics to react quickly.

Sentiment analysis can also help gauge the reaction to an advert that has been placed during the Super Bowl, meaning that future adverts can be changed based on feedback. However, the likelihood is that this would be a used prior to the Super Bowl slots or even variations and themes being tested through online video, as the average cost of a Super Bowl spot was $150,000 per second.

It is not only the product within the advert that is significant, with the brand itself often judged by their Super Bowl ad.

After the game has finished and the next day people are talking about the game, one of the biggest discussions online is not about the game itself, but about the adverts and the reaction to them.

This year for instance, WSJ are even running a poll for their readers to vote on which advert was the best at the Super Bowl. These adverts are seen by millions during the game itself and this can either see sales spike from a good ad or drop if the advert bombs. Examples of this from this year are likely to be Mophie and Budweiser, who both had positive sentiment around their ads, however, Chevrolet had a negative one to theirs, even having Anna Kendrick tweet that she didn’t like the ad to her 3.3 million followers.

We are seeing that in 2015 data is as much a part of all elements of the Super Bowl than the actual game itself. It is truly the grand final for companies looking to advertise, and much like the teams on the field, they need to prepare and train to get the result they want. 


Read next:

Maximizing the Organization: Turning Data Into a Competitive Advantage