Every year on Super Bowl Sunday, America watches for two things: great football and great advertising.
This year, 30-second spots cost a record $3 million. The usual suspects forked over the cash (Coke, Pepsi, Budweiser, GoDaddy.com, etc.), but there was a twist.
MillerCoors turned the practice of paying a few million for ad time into an economic statement and a highly-effective viral marketing campaign by creating a one-second ad for Miller High Life.
The brewing company launched a web site with “out takes” from the one-second ads that feature comedian Windell Middlebrooks. In the weeks leading up to the big game, the site grew in popularity.
On Super Bowl Sunday, those in the know watched commercial breaks without blinking to try and catch the ad.
The objective for this study was to use Spark to identify influential web sites and methods for spreading the ad campaign. Sites used in this case study specifically mention the Miller High Life one-second Super Bowl ad.
Data was collected from January 28, 2009 to February 2, 2009. Since Spark can look back in time, many sites in the study were published before January 28.
Category Breakdown
Each site in the study falls into a separate category, or type of site. A News site is considered a newspaper or television station web site, a Social Network would be a Twitter or MySpace page, while an example of a Social Bookmark would be digg.com. A Video site is YouTube, Vimeo, etc.
This is the breakdown of the categories found in this case study. The percentage next to each category represents the percentage of web sites in that category (rounded to the nearest one-tenth percent.)
• Reference: .3%
• News: 19.1%
• Forums: .5%
• Company/Organization (MillerCoors): .8%
• Blogs: 20.9%
• Videos: 1.8%
• Social Networks: 55.4%
• Social Bookmarks: 1.3%
The domination of Social Networks, News and Blogs over categories like Company/Organization and Reference suggest the campaign was successful in generating conversation about the ad and Miller High Life.
Sentiment
Sentiment is measured in seven degrees: Very Negative, Negative, Slightly Negative, Neutral, Slightly Positive, Positive and Very Positive.
It’s important to note that not every web site contains sentiment. Many news stories or press releases are not in these categories because they do not contain opinions.
In this case, sentiment toward the one-second ad campaign was overwhelmingly positive. Here are some key sentiment findings from the study:
• All three degrees of positive sentiment – Slightly Positive, Positive and Very Positive – totaled 92% of all sites with sentiment.
• Negative and Slightly Negative sentiment totaled 7.5%.
• Sorting the list of sites revealed that 67% of sentiment occurrences came from the social networking site Twitter.
• Another sort reveals that 73% of positive sentiment occurred before the ad aired on Super Bowl Sunday, meaning the campaign received large amounts of good publicity on anticipation and the video “out takes.”
• Negative sentiment totaled 19.2% of all sentiment occurring during and after the game, compared to 4% in the days leading up to the game. Most of these negative comments came from people who were not able to view the ad in their city.
Inbound/Outbound Links & Influential Sites
Spark lists the inbound and outbound connections from each site as well as an influence ranking for each page. The influence ranking is based on a third-party domain ranking system. A first-order connection means a direct link between two sites (think “six degrees of separation”).
Here are key findings from the inbound/outbound connections and influence rankings:
• 33% of all sites in the study have first-order inbound links to MillerCoors site www.1secondad.com
• Based on inbound connections, the Milwaukee Business Journal and MSNBC.com are the most influential news web sites.
• Digg.com was the most influential domain based on the ranking system.
Semantics
Spark uses a Semantic Cloud to list the words that appear most frequently on a set of web sites. These can be sorted by category, too.
The default cloud shows the top 100 most frequent words across all sites in the MillerCoors case study. The words in large font are the most frequently used.
Some words that stand out in this specific cloud are superlatives like brilliant, genius, good, great and love. Further investigation revealed that these words were used by the audience to describe the ad campaign.
These five words appear on 50% of the sites in this study.
Another word worth mentioning on this specific view is RT. It looks like nothing important to most people, but in Twitter-speak this stands for “Re-Tweet” or “pass along this message.”
This means those using Twitter to discuss the ad were passing along information from other sources so all of their “followers” could see it. Re-Tweeting is a powerful behavior that allows information to go viral in a short amount of time.
Only 25 sites in the study use the RT method, but if each of those users had 100 followers that is an extra 2,500 people who see the information. A RT looks like this:
Virtualization
This is the embedded file of the Virtualization, or visual map, of the one-second ad’s presence online. Here are a few keys to understanding the Virtualization:
• Each sphere is a web site mentioning the one-second ad campaign.
• The lines connecting spheres represent sites that link to each other.
• Green spheres are web sites with positive sentiment.
• Red spheres have negative sentiment.
• Gray spheres are neutral or no sentiment.
It’s the first week of the month, which means it’s time for the social media minds of Kansas City to meet for breakfast.
The Semantic Cloud provided critical information about communication at SXSW for PBS. The cloud lists the top 100 most frequently used words on all of the sites in an ecosystem (dismissing “a,” “the,” “and,” etc.) in the form of an organized tag cloud.
