Last night was reportedly the most watched presidential debate ever. Some are estimating over 100- million viewers watched Hillary Clinton and Donald Trump go at it. I set up some Dynatrace Synthetics tests to compare the sites that would be live streaming the debate. I monitored from nine locations across the US, every five minutes, using both a Chrome and IE browser. I divided the tests into two categories Traditional Media and New Media. The traditional media sites included ABC, CBS, MSNBC, Fox, CNN and PBS. The new media sites included Facebook, Twitter, YouTube and AltSpaceVR.
Our tests were directed at the specific live-streaming pages for the debate (ie. Not the home or landing pages of the sites). All of the sites provided high success rates at being available to end users and providing quality streams for the duration of the event.
The above view shows how Dynatrace can quickly create customized dashboards to look at specific categories, tests, data sets, etc. I divided the tests into categories and, in general, the traditional media sites were all significantly slower than the new media sites. What was also interesting was that as the debate started the performance of some of the new media sites began to slow down, but still not as slow as the traditional media sites.
Looking into some of the detail helps explain why the slowdown occurred. When we compare some of the Key Delivery Indicators (these are physical aspects of how the pages are delivered) to response time we see some interesting correlations. Traditional media sites tend to be much heavier than new media sites (the pages have higher byte counts meaning they are delivering more content). Traditional media sites also tend to be more complex (the pages have more object and connection requests) than new media sites. Finally, traditional media sites use many more third parties (ad providers, analytics, trackers, social media, etc.) than new media sites.
How the pages are delivered can be just as important as the technology stack being used to deliver them. What does this mean? Traditional media sites need to step up their game when it comes to delivery performance if they want to attract more millennial viewership. Millennials as we know have little patience for waiting for pages to load and will quickly abandon sites in favour of something faster.
Now that we’ve looked at the streaming media sites performance, how did the candidate sites perform during last night’s debate?
(Note: The following data is observationally empirical and there no attempt to insinuate any political bias either way in the following data)
Below is a view comparing the performance (load time) of Donald Trump’s home page compared to Hillary Clinton’s. During the debate the Trump site did see a small performance event which caused page requests to slow down and, in some cases, we saw some 503 errors indicating a potential server side issues.
In comparison, Hillary Clinton’s site saw no change in performance, even after she invited the audience to go to her site.
We can look at the past six months comparing the performance of the candidate’s websites and we can see that around the middle of July, Hillary Clinton’s website changed, and that performance since that time has been very consistent and fast.
Last night’s debate was one for the record books. Possibly the most watched political event of our time, we can see that the approach to delivering this content differs considerably between traditional and new media. As Millennials become the target demographic, how content is being delivered to them must be carefully considered. Performance is a key element of this effort, and the candidate’s digital teams need to consider it. Hillary Clinton on several occasions called for the audience to go to her website. I find this very interesting as it shows how the concept of Digital Disruption is permeating every aspect of modern society, including politics.
Keep watching as Dynatrace reports on the other debates, media and candidate sites leading into final weeks of the election.