Top 5 data trends that shaped mobile app marketing in 2020
‘The coronavirus pandemic’ is the first set of words in AppsFlyer’s 4th annual top 5 trends of the year post, and for good reason.
The pandemic has led to more home time and unfortunately more alone time, as people all over the world took to their mobile devices to meet a wide variety of needs: emotional (gaming, fitness, social networking, streaming, shopping…) and functional (finance, business, food delivery, and once again shopping…).
We’ve seen the shift to digital in recent years, but then came COVID-19 and greatly accelerated existing trends. Among others, the adoption of mobile apps and the understanding among brands that this channel should be at the forefront of their presence.
2020 was also the year Apple introduced dramatic privacy-related changes as part of its iOS 14 release. The announcement sent shockwaves across the industry, prompting Apple to postpone its enforcement to “early 2021”. What impact did the impending shake up have on 2020 campaigns? Find out ahead.
Our data sample covers 48 billion installs in 2020 of over 30,000 apps with a minimum of 5,000 installs per month.
Click to explore (and back in browser to go back):
1) App downloads climbed 33% as UA spend hit $74.6 billion
Mobile app marketers spent $74.6 billion globally in 2020 to drive users to install mobile apps. The figure is slightly less than the $76.2 billion we predicted in February, but still represents a significant 30% increase compared to 2019.
A platform breakdown has Android at $48.5 billion and iOS at $26 billion. Apple’s OS commands 35% of spend despite its 18% market share of non-organic installs because its cost per install is more than double that of Android.
Despite the surge in demand for mobile apps, the high cost of media prevented marketers from further scaling, as budgets continued to shift to remarketing (see trend #2 for more).
The number of app downloads increased by 33% in 2020, up 25% compared to 2019. We can see that both pure consumer intent and marketing efforts contributed to this leap.
But a closer look shows opposite trends, despite the impressive growth on both ends. While the growth rate of marketing-driven installs (non-organic installs or NOI) dropped by 27% this year, pure consumer intent — or organic installs — jumped by almost 60%.
We can also see that remarketing adoption continues to rise. This trend is driven by 1) the significant increase in the cost of installs, compared to the far more affordable price of remarketing media, and 2) the focus on cultivating loyalty amid retention struggles.
In 2019, remarketing exploded among mobile apps, so keeping up with that pace is unrealistic, in spite of the above. Although growth was cut by half in 2020, it remained extremely robust at 82%.
A deep dive into this unusual year demonstrates just how significant the March-April lockdown period was, especially for marketers, with double the growth rate compared to Aug-Oct (in other words, NOI grew 57% in March-April 2020 compared to March-April 2019, while growth in Aug-Oct 2020 increased by “only” 29%).
2) Cost per install spiked 30% post-lockdown leading to higher spend despite drop in NOI
Another view of 2020 trends showed just how much the cost of media impacted marketing decisions.
Early lockdowns in March-April led marketers to aggressively promote their apps amid extremely high demand, particularly among the quick-to-react gaming apps. The fact that CPI was at a low point — with traditional big brands holding back on marketing amid the uncertainty at the time — helped apps scale.
But starting in May as lockdowns were eased in many markets and anxiety lessened to some extent, big brands returned. Furthermore, many other brands who had started shifting to digital, realized that they needed to move much faster. As previously mentioned, the pandemic was an accelerator of many processes.
The end result: the competition for user attention accelerated and with it cost per install jumped 30% between April and November — a trend that was seen in both SRNs (self-reporting networks) and non-SRNs. Many app marketers tend to look at the app ecosystem, and even their own category’s competitors. But the reality is that they compete with anyone who advertises on a given platform.
Indeed, the number of non-organic installs also dropped as the cost of media started to climb (there was also a decline in pure consumer intent after heavy activity during lockdown as the trend line in organic is similar: dropping but not as much as NOI). In total, NOI was down 22% from the April peak to November.
As UA dropped, remarketing consistently climbed with lower prices and a large number of new users acquired early on who could be remarketed. All in all, the number of app remarketing conversions (counted as users who open an app after clicking on a remarketing promotion) leaped 70% from March to November.
Another indication of just how significant the cost increase was can be seen in the UA spend trend. Despite the decline in NOI, budgets increased by 25% in Oct-Nov (compared to June). In most of H2, app marketers were spending more money and getting fewer users.
3) App revenue on the rise across revenue streams: IAP, IAA, and subscriptions
The massive boost in the number of new app users in 2020 has led to an uplift across revenue streams.
When it comes to the largest revenue stream — in-app purchases (IAP) — we can see that despite overall revenue growth, gaming and non-gaming showed different patterns early on in lockdown (and therefore were presented separately).
Non-gaming apps, mainly the large traditional brands, reduced or ceased marketing spend amid the uncertainty; also, most people were only looking to purchase essential goods during the first lockdown (in some countries, only businesses that sold essential goods were open). After lockdown, the new normal set in and that was when revenue of non-gaming apps in multiple categories started to climb. By November, it had surged 35%.
Gaming IAPs, on the other hand, started to rise early as gaming app marketers employed aggressive UA strategies during lockdown leading many new players to install gaming apps. Games were then able to monetize players who remained loyal in the following months through in-app purchases.
In-app advertising (IAA) for gaming apps (the sample size was not statistically significant for non-gaming apps) also made strides throughout the year, peaking during lockdown, and then climbing 30% since July.
Ad revenue is mostly related to CPI and app sessions. Despite the drop in the number of sessions since the lockdown surge, the rise in CPI increased at a greater rate, and with it CPMs for publishers.
Subscription revenue started to climb in March and by April had already jumped by 40% compared to February. Clearly, while at home, more people streamed more music and video content, while also subscribing to other services like Health & Fitness, Education and even Dating apps. In fact, the average app running subscriptions increased its revenue in 2020 by 56%, with the large players almost doubling their income.
4) Massive organic demand for non-gaming apps, while gaming apps surged with UA
A comparison of the different app categories, sorted by year-over-year growth of organic installs also tells the story of the pandemic.
People around the world used apps to socialize (Social) while physically away from each other, used various tools and utilities while working from home (Business, Utilities), shopped, watched videos (Video Players, Entertainment), worked out and meditated (Health & Fitness), educate (when school was out) and of course play games. Regrettably, there was negative demand for Travel apps, although it has picked up in the last couple of months.
We can see that some verticals were looking to capture heightened consumer intent and doubled down on spend (e.g. Education, Gaming and Video Players), while others grew largely due to organic demand (e.g. Lifestyle, Health & Fitness, Shopping) that was not matched by marketing spend (e.g. Lifestyle, Health & Fitness, Shopping, Utilities) — perhaps a missed growth opportunity, or a desire to minimize cannibalization of organic traffic.
A look at the top growing markets by non-organic installs — sorted among the largest 30 markets by number of total installs — shows significant growth opportunities across the globe: from the Middle East, Asia, Africa and Latin America.
Markets with significantly higher organic than non-organic growth are a potential growth opportunity: these include Japan, Philippines, Iraq, Saudi Arabia and Indonesia. Even the world’s most popular app market, the United States, had 75% higher organic growth.
5) Almost 1 in 10 non-gaming installs had a preceding visit to the brand’s website
The reality is that, in 2020, there still remains a significant divide between user behavior and marketing measurement.
On one hand, there are the end users whose conversion journeys are becoming increasingly complex, involving multiple devices and touchpoints. On the other hand, there are the marketers who measure user actions across different channels and devices, but do so in silo from one another.
But guess what, an analysis of 45 apps (mainly eCommerce, Finance, Media & Entertainment, and Food & Drink — all of which were live throughout the measured time frame) found that the number of app installs with a preceding visit to the brand’s website has almost doubled this year. In fact, in almost 1 in 10 installs, the web was part of the user journey.
This is big news for marketers. It means that they can also acquire users on the web where the cost of media is cheaper, and drive them to the app via owned media promotions (website button, email etc.)
BONUS! Privacy measures in iOS 14 have yet to impact campaigns (focus on yet…)
In June, Apple rattled the app marketing ecosystem with its announcement of a new privacy-driven mechanism to be implemented as part of iOS 14. The AppTrackingTransparency framework (ATT) required users to actively opt-in to IDFA collection when using the app
In practice, such a mechanism would de-facto eliminate the ability to use the IDFA for measurement and attribution. In October, we found that 99% of users would NOT opt-in when asked to be tracked.
Apple’s decision to postpone the required change to “early” 2021 led to a collective sigh of relief.
But while things shook under the surface, the fact that apps were not required to implement the changes allowed marketers to continue as usual. Indeed, the data below shows that, for the time being, the upcoming release has yet to shift budgeting decisions — neither for UA, nor for remarketing.
Things will look different in the ‘Top 5 trends of 2021’ post. After all, advertisers spent $26 billion on iOS UA alone in 2020 (see trend #1 for more).
The million dollar question: What will 2021 look like?
The truth is that predictions are never easy, but 2021 is truly a difficult one. COVID-19 will likely continue to impact social behavior at least in the coming months even if a vaccine is hopefully approved and distribution begins. That means continued economic uncertainty for many households and businesses, which can affect consumer spend.
Having said that, we’ve seen the rise of digital during social distancing, which continues to present an economic opportunity for mobile apps. Even if this crisis will [hopefully!] come to an end in 2021, digital acceleration is a fact as scores of people across the globe have already adopted mobile apps in 2020 as part of their daily routine.
Beyond the economic conditions, the expected release of the ATT framework in iOS 14 will have a major effect on the app marketing space, and on measurement in particular.
In addition to several significant unknowns that Apple has yet to finalize or clarify, marketers also face many uncertainties: how will measurement actually work, what will be the impact on the cost of media, what will the media mix look like, what will be the future of remarketing (which we’ve seen surge in the last couple of years), how will machine learning-driven ad platforms adapt, to name just a few.
In the absence of IDFA, more focus will be placed on various other forms of measurement, especially aggregated measurement, in addition to incrementality, probabilistic modelling, and the expansion of web-to-app attribution models.
Ultimately, change drives innovation. Measurement will adapt to a new privacy-centric reality that eliminates the bad of personalized advertising and keeps the good to allow data-driven performance marketing to continue to thrive.