Where Has All The Time Gone?
I’ve managed to get a sneak peek into how much time was spent on social platforms in Q3 2022 from Evercore ISI. The rest are my thoughts on the topic.
One way of measuring user engagement – assessing the average time a typical user spends on a given platform. (Pretty much everyone now is using iOS time in the app obtained either via data.ai or SensorTower as a suitable proxy metric for the US population; this methodology doesn’t work nearly as well outside the US.) Using time on a platform as a proxy for user engagement is yet another approximation and simplification, since engagement comes in many flavours, including, but not limited to, to the number of content items consumed, created, shared, commented on, etc. A single metric of minutes per day hides all the nuance associated with user behaviour. So, let’s be clear about the core assumption: a higher time on the platform is (generally) better than lower time on the platform as it (again) is considered a positively correlated factor to service monetisation.
When it comes to DAU (daily average users), this becomes more curious: steady DAUs mean that the audience simply replicates itself (where new users roughly equal the churn), or (!) that a company has decided to fight bots, which exaggerated operational numbers. While it’s painful to admit that some growth comes from bots, but it’s very brave and necessary to trim the fat and restore accuracy in the numbers where applicable. (I don’t specifically mean Twitter, by the way, as it’s a very easy target to pick.) Similarly, unexplained DAU growth with no matching increase in engagement is a sure sign of either a paid marketing campaign with a much larger budget than usual (and - since it’s not fully organic - with lower user retention), or .. you guessed it … a malicious actor who found a way to mass-create accounts and imitate activity.
Sadly, iOS 14.5 has made it much less profitable to attract iOS users due to severe limits to how they can be targeted and monetised, so if in the past it was possible to quickly discover anomalies in the iOS/Android acquisition/usage split, nowadays this signal is all but meaningless. Further, the proliferation of Android emulators coupled with proxy networks has led platforms to invest time and efforts into determining which users are human versus who are not monetisable.
TikTok hasn’t eaten up into the Facebook’s/Instagram’s territory just yet – at least, judging by the marginally growing share of time users spent on Meta platform apps compared to 2019 (what I consider the last sane year for platform apps). It’s incredible that TikTok manages to coax an average user into spending 93 minutes per day glued to this app (+5% YoY).
YouTube continues to hold its ground steadily and is increasing its share in the US TV viewing time. Shorts have at least some impact. 72 mins per day in Q3 2022 is excellent.
Pinterest is slightly down (12 mins daily usage vs 13 mins pre-Covid). Looking at their revenues, the argument that more time in app leads to proportionally better monetisation is generally upheld.
Snapchat’s usage has dropped from 27 mins / day (2019) to 20 mins (Q3 2022). In spite of this, Snap’s DAU is up 2% YoY.
One may wonder if comparing messaging apps, social networks and video-centered apps is really comparing apples to apples or whether this is an attempt to establish some common ground for comparing large private (TikTok) and public (everyone else) companies. But it goes without saying that the ARPDAU (average revenue per daily average user) varies significantly between the platforms of different kinds.
P.S. As you probably know, our company FunCorp has invested in Yepp, a company developing an entertainment app focused on quality memes and funny content. We’re currently inviting entertainment content creators to monetise their content via our app, and have an affiliate program for referrals. You can read more here.
(Yes, I’m still working on the best sales pitch for the product, but it’s the real deal and I’m also contactable on my email or LinkedIn).