Super Bowl Web Traffic Blitzes Advertisers
Super Bowl Web traffic crushed sites from Coca-Cola, Doritos and other brands that blended interactive marketing with televised commercials. Online statistics show that site design proved just as important as network capacity.
February 7, 2013
![Network Computing logo Network Computing logo](https://eu-images.contentstack.com/v3/assets/bltde8121fc52c5c8f3/blt3f3d0318f746b1c2/65a530e4187606040a1d8b8c/placeholder.png?width=1280&auto=webp&quality=95&format=jpg&disable=upscale)
The Super Bowl is more than a competition to determine the best football team in the NFL. It's a testing ground for the best advertising and consumer marketing U.S. advertisers can throw at the estimated 108 million Americans watching the game.
"[The Super Bowl] has become as much a digital event as a marketing event," said M.J. Johnson, director of product marketing at Akamai, one of two content-delivery networks (CDNs) that helped spread the load of responding to user requests to servers around the world.
"From a traffic perspective it grows every year, and grows in interesting ways," Johnson said. "This year it wasn't just incremental traffic, it was a concentrated effort to have the online or streaming experience supplement the TV experience."
Akamai measured Internet traffic during the Super Bowl. Source: Akamai
That was a big change from the 2012 game, according to ReadWriteWeb, which argued that advertises failed to "meaningfully connect their message to their social media platforms." However, the effort carries some risk--including high expectations among viewers for interactivity, streaming media and quick response times.
Slow response times, more than almost any other factor, can ruin a user's view of a brand and a website, even for sites that perform well outside of game day, according to Ben Grubin, marketing director at Compuware APM, an application performance management company. Compuware used its performance-monitoring tools to keep tabs on the performance of advertisers' Web properties during the Super Bowl.
"All the advertisers had obviously prepared for higher traffic; almost all of them were using CDNs to help handle spikes," Grubin said. "Response times were only partly dependent on the network. Times varied more depending on who relied on third-party services with Facebook Like buttons and other social network tie-ins, and how heavy the site itself was. The ones with lots of scripts to drive the interoperability, a lot of objects to load before the page displays, and heavy-duty graphics didn't do so well," Grubin said. "That combination can be deadly."
[ Join us at Interop Las Vegas for access to 125+ IT sessions and 300+ exhibiting companies. Register today! ]
Most specifically, sites for Doritos, Coca Cola and Universal Studios were consistently the worst among advertisers, who paid an average of $4 million for every 30-second TV spot.
Compuware's Web performance monitoring shows the three best- and three worst-performing websites during the Super Bowl based on response times. Source: Compuware APM
"Coke's average response time was about 10 seconds, but that includes an hour before the game and an hour after," Grubin said. "Right after the ad aired, response times blasted to between 20 and 35 seconds, and some up to a minute. People won't wait that long for a response."
It can also cause companies' sites to lose out on serendipitous opportunities.
"There was a long blackout during the game, which some companies used as an opportunity to drive more traffic to their sites," Johnson said. "And many connected users evidently had some idle time in the middle of the broadcast to go on Facebook and Twitter and talk about the game and the ads or whatever."
Next Page: Coke Popular But SlowCoke's ad featured three teams "lost" in the desert and gave viewers the chance to vote for which should find the secret cache of Cokes. After the game, Coke ran the commercial's coda: another ad revealing which team got the soda (the showgirls).
"Coke kind of stuck its neck out with a very social campaign," Grubin said. "They had people come in and vote and it was very interactive, so there were a lot of scripts and links to Facebook and other third parties, and a lot of objects that had to be downloaded and a very high density of graphics. It was a beautiful site; if you looked at it before the game it worked well."
However, response times of 20 seconds to 30 seconds during the hot times right after Coke's ad aired drove away many customers; others never got a reply at all.
"Some of our testing nodes came back with no data because, if they don't hear back 60 seconds after a request, they assume it failed," Compuware's Grubin said.
The voting function that was the core of the campaign was down periodically during the game, as were video features, according to Mashable. Nevertheless, the campaign generated more than 11 million "fan engagements" and more than 86,000 mentions of the campaign on Twitter and other networks, Coke told the New York Times.
In total, the football game and the ads in it generated more than 24 million online conversations during the game, according to Networked Insights.
The brands that had the quickest-responding numbers had sites were far lighter than the bottom three, and minimized the number of third-party hosts that had to be tied in to home page requests.
Go Daddy managed a response time of 0.873 seconds for a site with 10 hosts and 35 objects. It also apparently didn't excite viewers enough with its ad to send them to its site, though the ad--which featured a supermodel making out with a geek--was among the most-discussed during and after the game, but didn't draw huge in-game traffic spikes.
Paramount Studios was the only other site to average response times under a second, at 0.97 seconds, with just three host contacts and 41 light-ish objects required to display the page.
"It really comes down to site design and what you're making the user wait for when the page loads," Grubin said. "Almost all the advertisers used CDNs to help spread the load and had probably increased their own capacity, too, so just the speed of the network or number of requests wasn't the thing that made the difference."
About the Author
You May Also Like
Maximizing cloud potential: Building and operating an effective Cloud Center of Excellence (CCoE)
September 10, 2024Radical Automation of ITSM
September 19, 2024Maximizing Manufacturing Efficiency with Real-Time Production Monitoring
September 25, 2024