Did the first U.S. Presidential Debate Help Twitter?
According to CNN, the first presidential debate of 2016 was the most-watched debate in history, with a record-setting 84 million people tuning in to watch Hillary Clinton and Donald Trump. But what do we know about the amount of activity on Twitter during the debate, especially in key swing states?
Twitter has played a central role in both Hillary Clinton and Donald Trump’s communication strategies, and each candidate arguably employs it to different effect. Trump has infamously used Twitter as a personal bullhorn for his famously unscripted ideas, and both candidates have employed the social media network as a public sparring ground.
This year, Twitter has proven to be an excellent way for political pundits and spectators alike to watch this year’s race unfold. And despite a troubled year and rumors that it may be sold to another tech or digital publishing giant, Twitter has stepped up to its role: it’s livestreaming all three presidential debates (including the initial debate in New York).
Verto Analytics looked at Twitter activity across the U.S. and in key swing states during the first presidential debate on September 26 to determine if the debate had a measurable impact on the social network (among American adults, ages 18+), and what this could mean for Twitter and the two candidates as we head into the second debate scheduled for this weekend. Here’s what we discovered.
A Distinct Spike in Activity on Twitter
According to Verto Analytics data, Twitter experienced a distinct spike in traffic from across the U.S. on September 26 – and in some cases, this heightened traffic continued through the rest of the week. As shown in the chart below, the greatest increase was seen in external traffic, and not an increase in users on Twitter’s website or flagship app.
The Importance of External Traffic
Many apps, services, and websites generate traffic in multiple ways, but only a portion of this traffic should be actually attributed to active users (or even registered users). As we see in the graph above, external traffic is responsible for more that 80 million unique users on September 26 alone – that’s nearly a 30% jump in traffic from the previous day. However, external traffic is a broad category: most importantly, it includes people who are exposed to Twitter content with being logged in to the site. External traffic also includes users who are exposed to Twitter content on sites that are not part of Twitter’s own domain or apps (eg., feeds embedded on sites and services, such as news sites that feature a scrolling Twitter feed in the sidebar). And, in the case of the presidential debate, external traffic could also include any user who watched the livestream of the presidential debate without logging in to Twitter.
As we also see in the graph above, Twitter users who logged in to either the Twitter app or website represented a far smaller group of Twitter users overall, and the jump in user activity for these two properties on September 26 was far less dramatic that the spike we saw in external traffic.
Iowa, Nevada, and Florida: Key Swing States on Twitter
Verto Analytics data shows that while residents in nearly all eleven swing states contributed to a jump in Twitter activity during the presidential debate, some states were responsible for truly dramatic increases:
External traffic from Florida, Iowa, Nevada, North Carolina, and Virginia all created enormous user spikes for Twitter on September 26.
Iowa and Nevada were particularly active; between September 25 and 26, Twitter’s external traffic from Iowa jumped 104%, to nearly 720,000 users. According to 2015 estimates from the U.S. Census Bureau, that’s about 30% of Iowa’s total population of adults over 18 – which is also the federal voting age in the U.S. The impact in Nevada was even more profound: Twitter’s external traffic from the Silver State jumped 106% between September 25 and 26, representing almost 1.1 million users, or 48% of the state’s adult population. (To put this in perspective, only 23% of the entire U.S. population uses Twitter on a regular basis, according to the latest estimate from the Pew Research Center).
Meanwhile, Florida, an extremely influential swing state, sent approximately 7.9 million external visitors to Twitter on September 26. While the jump in user activity was relatively small compared to Iowa and Nevada’s numbers, nearly half (49%) of Florida’s entire adult population had some form of exposure to Twitter on the day of the presidential debate. Regardless, for these three states, each of which could swing the election by a few thousand votes, huge portions of the voting population interacted with Twitter at some point on September 26 – and perhaps spent enough time online to reach a new decision about their preferred candidate.
Surprisingly, Twitter traffic – both from external sources as well as from logged-in users – in Ohio, another key swing state, was relatively unremarkable over the course of September 26. External traffic from Ohio actually dropped by 8% on September 26 (compared to the previous day), while users on twitter.com climbed a relatively modest 19%. According to the New York Times, no presidential candidate has won without carrying Ohio since Kennedy’s 1960 race, but the state has slowly been losing its influence in this year’s election. The relative apathy reflected on Twitter may be a symptom (or cause) or this electoral sea change.
Is the U.S. Presidential Election Season Good for Twitter?
Not all Twitter traffic on September 26 – external or otherwise – can solely be attributed to the presidential debate. But the undeniable spike in user activity does indicate that the debate had some impact on these numbers.
The value of social media when it comes to the U.S. presidential race is no longer a question – it’s a given. President Obama has been hailed as “the first president of the social media age.” And according to the Pew Research Center, “roughly a quarter of U.S. adults (24%) turn to social media posts from either the Hillary Clinton or Donald Trump campaigns as a way of keeping up with the election; this is a higher share than say they turn to the candidates’ campaign websites (10%) or emails (9%).”
While the election may be great for Twitter’s traffic numbers, however, it’s not a great indicator of user engagement or even new user acquisition. Even the massive external traffic numbers aren’t necessarily good news for Twitter, which has been suffering from stagnating user growth and a tumultuous executive lineup. The majority of Twitter’s traffic comes from external sources – that was especially true on the day of the presidential debate, but it’s also a consistent trend otherwise. And while external traffic makes for some impressive numbers, they’re not necessarily the type of users who stick around and engage with your product – or with your advertisers.
For example, there are 211 million unique users who are exposed to some form of Twitter content every single month in the U.S. But if we calculate the number of unique users on Twitter’s own flagship website, across all digital devices, then the total number of monthly users is closer to 97 million. That is still a notable figure, but it’s much lower that the total reach of 211 million users. And if we look at Twitter’s mobile app, our data reveals there are actually only 30 million unique users a month (and 9 million unique users daily). While we saw some incredible jumps in external user traffic (a 106% increase in Nevada, for instance), Twitter’s mobile app only saw a 7% increase overall.
This is also bad news for advertisers who rely on Twitter as a major distribution platform: external traffic can’t necessarily be tracked or targeted, and it’s unclear how much of this traffic is even exposed to Twitter’s ad products.
Although it’s reportedly on the verge of being purchased, Twitter (or its buyer) will need to develop a way to capture some of this external traffic and either monetize it via ads, or convert them into active users. Using the increased attention during election season may be a prime opportunity, if Twitter can react in time.