Age of Twitpocalypse

The Four Horsemen of Twitter’s Impending Doom… or Not.

Apparently, the world is ending. Since my post House of Twitter Cards was published, pundits have declared it’s the End of Days for Twitter. Slowing growth in user engagement and acquisition have spooked Wall Street analysts and technology pundits. While there is a legitimate cause for concern, the metrics raised as evidence of the coming apocalypse can be misleading. As I’ve discussed in the prior post, many (if not the majority of) Twitter impressions live outside of Twitter in the real world—on a live feed during NBA Commissioner Adam Silver’s press conference, on a CNBC newscast on Carl Icahn or during an episode of Glee. There is tremendous upside in these users and even more upside in the ARPU of existing and potential Twitter users. But if we are entering Twitter’s final days as the naysayers and soothsayers suggest, who are Twitter’s harbingers of doom, the alleged Four Horsemen of the Twitpocalpyse?

The White Horse: Slowing Engagement

Twitter engagement is troubling. Timeline Views per MAU in the last two quarters are down significantly from their peak in 2013. To offset this, monetization in the past two quarters is up. A cursory glance at these data points implies that while Twitter is growing more monetizable users, they’re not nearly as engaged as before. Why is this?

Source: SEC filings

Twitter follows the 90-9-1 rule. The vast majority of Twitter users and potential Twitter users will rarely or never tweet. There’s nothing wrong with that. Most people read blogs like this one; they don’t write them. Dick Costolo has mentioned Twitter is taking steps to make its user experience more friendly and the first step it has to take is to make it more reader-friendly.

Consider this use case: I’m watching the NBA Playoffs on TNT right now, and there are various hashtags being used like the generic #NBAPlayoffs, #GetWellSager and #Shaqtin. If I wanted to follow that content online and I’m not already a Twitter user, I have to endure an arduous on-boarding process on my mobile/tablet app before I can get to a screen where I can actually search for the NBA hashtags. Twitter probably has already lost me as a user.

Instead, what if the Twitter app gives a new user the option of bypassing the on-boarding process and takes them directly to the top trending stories on Twitter in one click? I suspect a large portion of users at a certain time and geography will be looking for the same 9-10 trending hashtags. For example, YouTube’s mobile/tablet app has a more welcoming UI. Here is the launch screen when I open the app for the first time.

Source: YouTube App

If I want to jump right into watching content, I “continue as guest” and it takes me immediately to some curated videos and a selection of categories to follow.

Source: YouTube App

Twitter could do something similar but with popular trending hashtags, celebrities and content creators.

To prevent the raw hashtag feed of #NBAPlayoffs from devolving into the inane noise of a Yahoo! News comments section, Twitter can also employ algorithmic filters to highlight the top tweets and remove bad language or hate speech. Like the Apple app store, it could also use human curators to promote top tweets from celebrities or hidden gems. If the user wants to join the conversation, she can go through a light-weight join process (name and password only). Since the app already knows what you like based on your reading choices, it could suggest accounts to follow. This reader-friendly UI could boost engagement amongst new Twitter users and perhaps even encourage user growth, another metric on the decline.

The Red Horse: Slowing User Growth

Last quarter, Twitter announced it had 255 million active users, up only 5.8% from the previous quarter and 25% from the previous year. A widely circulated chart shows the steep decline in year-over-year growth for Twitter user growth.

Source: SEC filings

Of course, there’s more to this. Here is the same chart with Facebook’s user growth but in the context of their respective IPO dates.

Source: SEC filings

While the drop for Twitter is slightly steeper than Facebook’s, the sky is not falling. This decline is expected for a company that is hitting declining returns in its first core market. For example, when Facebook went public in May 2012, its previous quarter’s q/q growth rate was 6.6% and y/y growth rate was 32.5%. Twitter’s is 6.4% and 38.9% respectively. Recall, Facebook stock wasn’t a Wall Street darling out the gates either. Facebook plummeted from its IPO price of $38 to $18 in mid 2012. It is now at $60. How did the stock rebound? The answer is mobile.

Facebook’s expansion onto the mobile platform saw its mobile revenue grow from 14% of Q3 2012 revenue to 59% of Q1 2014 revenue. While mobile revenue grew at a tremendous clip, mobile users in Facebook didn’t grow nearly as fast. In Q3 2012, 60% of all Facebook MAUs used a mobile device. In Q1 2014, it was 79%, only a 19% share jump. What changed was that Facebook was able to start monetizing its mobile users at a meaningful ARPU, growing it from $0.25 in Q2 2012 to $1.33 in Q1 2014. What does this mean for Twitter?

Twitter is already monetizing the >75% of its users already on mobile. While it’s far from maximizing its mobile revenue, mobile isn’t the hidden value for Twitter as it was for Facebook. As I’ve mentioned previously, most people Twitter without tweeting because they watch it on television or see it embedded in an article. These tweets are all monetizable impressions except they’re not being monetized yet. If this sounds like a familiar refrain, it’s because Facebook had the same issue with its mobile platform. Facebook introduced the first version of its mobile app in 2007; it took five years before it began monetizing it. How and when will Twitter begin monetizing its non-native impressions?

One clue is a recent interview Dick Costolo made on CNBC. Twitter’s CEO referenced how tweets about the Oscars had 3.3 billion impressions in the 48 hours after the Oscars compared to some popular YouTube channels with 3 billion impressions in a month. Many of those tweets were embedded into an article like this:


To monetize these 10s of millions of embedded tweets, you could insert a pre-roll advertisement like YouTube does. Or you can also add a mini sponsored tweet below the tweet. Here’s my shitty photoshop of what that could look like. I’m not a designer but you get the idea.

Source: I made this in Photoshop

How about television? With so many tweets referenced on news segments and during shows, how will Twitter monetize TV? Its recent acquisitions may shed some light. Four of their most recent acquisitions are on social television analytics: Mesagraph, SecondSync, Trendrr and Bluefin Labs. By better understanding the relationship between television and social, can Twitter convince television networks and content producers to engage more?

The Black Horse: Twitter Does Not Impact TV

Twitter and TV should make sense. After all, television shows routinely showcase the Twitter handle to encourage social activity. Some recent studies lead some credence to this hypothesis. According to Fox, Twitter and the Advertising Research Foundation, 92% of Twitter users have taken immediate action around a given show based solely on what’s in their feed. Another study by Symphony Advanced Media found women using Twitter while they watched the Oscars viewed the show 39 percent longer than women not using Twitter (83 minutes vs. 60 minutes). Nielsen even has a weekly Twitter TV top 10, like the one listed below.


However, a recent Financial Times article on NBCUniversal’s head of research Alan Wurtzel may have poked a hole in Twitter’s claim that social activity encourages increased television viewership. During the Sochi Games, Wurtzel and NBCU monitored social activity on the Olympics and predicted it would have an impact on viewership. Instead, they found only 19% of Olympic viewers posted on social media. There are legitimate caveats that the Olympics audience is too large to see incremental increases from social activity. Nevertheless, it does cause some concern for Twitter’s TV hopes.

While everyone is referencing the more quote-worthy phrase “the emperor wears no clothes” when referencing Twitter’s impact on TV, the part of the article that interests me most is another observation:

“NBCU had expected social media to have a dominating effect on viewership for the Games. However, during the 18-day period of coverage, just 19 per cent of Olympic viewers posted about the games on social media, the broadcaster found. Mr Wurtzel said that a show’s ratings are more likely to drive activity on social media rather than vice versa.

If increased television viewership could have a bigger impact on social media activity, could Twitter borrow the monetization model from YouTube and do revenue-sharing with content creators like studios, networks and celebrities?

Currently, YouTube has a 55-45 share with its content creators. YouTube ad units are videos or banners, which are largely driven by impressions. Because Twitter lives mostly on mobile, its ads can be CPM or CPA driven through the use of Twitter cards. This distinction is important not only because CPA ads typically have a higher price point, they also open up new opportunities for the content creators. Imagine searching for #NBAPlayoffs on Twitter and seeing tweets from @NBAonTNT or @ESPNNBA. Both networks can push app downloads into the feed for TNT’s Team Stream and ESPN’s SportsCenter apps, or sell that inventory to brands like Nike to promote its app, Training Club. With app downloads at $2-3 per install, this could become a meaningful growth channel. Content creators would be further incentivized to drive traffic to their Twitter accounts from TV and vice versa.

While a revenue-share may seem like a foreign idea for Twitter now, because of its acquisition of MoPub, Twitter is already in the merchant business.

The Pale Horse: Twitter is Not Mainstream

Amongst the punditry on Twitter’s recent quarterly announcement is a theme that Twitter is not “mainstream.” Explicitly, this may mean Twitter is not user-friendly enough for folks like my mom and dad, or even some ad executives. Implicitly, Wall Street is asking if Twitter’s Total Addressable Market (TAM) is large enough. I think most of us would agree that Twitter is ubiquitous in the sense you see tweets on TV shows or newscasts even when you don’t use Twitter itself. However, it’s possible none of the monetization suggestions I’ve made in this or previous post could see the light of day. In other words, what if this ubiquity isn’t valuable? This is why the MoPub transaction becomes more meaningful.


In a recent interview with CNBC, Twitter’s largest investor and former Googler, Chris Sacca, compared the MoPub transaction to Google’s $12.7 billion AdSense business. AdSense, unlike Google AdWords, sells ads that are inserted contextually in other publishers’ content. MoPub is the mobile ad platform Twitter acquired in last September for $350 million. According to Twitter, MoPub reaches a billion users on iOS and Android, effectively making Twitter’s advertising reach far greater than the 255 million active users on Twitter today. With the more recent announcement of native ads and app downloads on MoPub, the argument for MoPub as mobile app AdSense is compelling.

Google’s ad-serving business isn’t measured on a per user basis. If it were, it’d be the most valuable with over 1 billion active “searchers” a month. That’s a value of over $300 of market cap per user, which towers over Facebook at $120 per user, LinkedIn at $60 and Twitter at $88. But we agree that a per-user metric for Google is silly because Google is more transactional. So what about Twitter? Does it make sense to measure Twitter on its 255 million MAUs when their tweets are viewed by billions more?

Dick Costolo recently unveiled a new tagline for Twitter:

“We think of Twitter as a companion experience to what’s happening in your world.”

During the press conference and in subsequent interviews, Costolo repeats this tagline. When pressed, he admits in a Bloomberg interview that it’s not only for branding purposes; it also helps guide Twitter employees on the direction of the product. What does that mean from a practical perspective? Throughout my recent posts, I’ve remarked multiple times how Twitter is being consumed constantly by non Twitter users in non-native media like TV or embedded in articles—and much of it on mobile. The MoPub platform further extends Twitter’s advertising reach to app content being consumed by millions more of non-Twitter users. Through these different properties, Twitter has a proprietary real-time view into what is happening around you at all times. The progressive verb tense “happening” is deliberate—Twitter’s view into your world is past, present and ongoing into the future.

Google has (in)famously downplayed social signals in search, citing social’s ephemeral nature as unreliable. Others suggest it’s because Google doesn’t own the data. If Google doesn’t want this responsibility, could Twitter step in? It does, unlike Google, actually own the data. If Twitter can serve ads to you wherever you are, whether or not you’re a Twitter user or not, based on the context of what is going on around you, can Twitter become some version of real-life Google? I would say this would make Twitter quite mainstream.

The Second Coming

Have we all been looking at Twitter the wrong way? The whole time, everyone has compared Twitter to Facebook because they are both social networks started around the same time. However, whereas Facebook maps your social graph (connecting you to your friends and family), Twitter maps your interest graph (connecting you to your personal and professional interests). Facebook sits behind a walled garden of privacy while Twitter is typically used openly. Is it possible that Google, and not Facebook, is a more appropriate competitor?

It may be entirely a coincidence that in the last 48 hours, the companies referenced by Dick Costolo (Twitter’s CEO) and Chris Sacca (Twitter’s largest investor) are YouTube and AdSense, two Google products. Or if you want to share this tin-foil hat with me, let’s play a scenario that is likely happening today:

You’re on the street waiting for your Uber (you elitist San Franciscan!) and reading an article on your New York Times app that has an embedded tweet. You receive an email that someone has favorited a tweet you made earlier in the day. When your Uber arrives, you jump in and start playing a mobile game you just downloaded. It’s Free-to-Play so you click off the ad that’s served to you before you start crushing candies or something. All that time, Twitter has collected information about you that probably no other company can. It knows what you read, what you play, where you are, what device you’re using and it also knows those things for the people around you at any given moment.

I think we’ve all been a bit too hasty to declare End of Days for Twitter. What these revelations may tell us is that Twitter’s not done. If we’re going to stick with the theme, then the Apocalypse is probably the wrong analog. The Second Coming may be more appropriate.


  • Apologies to those offended by the religious references. Once upon a time, I taught Sunday School and watched a lot of X-Men cartoons.
  • If you liked this, follow me on Twitter: I tweet stuff on tech, startups, inbound marketing, growth hacking and Liverpool FC. My Medium blog posts are on startups and tech (which are linked below).
Show your support

Clapping shows how much you appreciated David P. Cheng’s story.