Ad channel diversification for a scaling mobile app in the user privacy era

AppTrackingTransparency (ATT) was a pivotal moment for mobile performance advertising. Deterministic targeting and measurement data, the pillars that advertisers relied upon to simplify scaling their app user base, were shaken. Perceived performance on core app advertising channels had noticeably dipped. And the biggest victim in all of this? Mobile apps who had yet to reach scale.

The large apps coped better—they found new ways to measure and target with their deep bench of resources, meanwhile ad channels whose performance suffered learnt how to use first party data to target and build models to overcome measurement challenges. The smaller apps, however, faced these challenges:

Ad monetization became more difficult as huge swathes of targeting data was lost.
User acquisition performance became more complex and harder to comprehend as many advertisers did not have deeply sophisticated data science and engineering resources to do this.
At a smaller scale, the impact of the privacy threshold on SKAdNetwork (SKAN) was bigger for these growing advertisers, making them fly even more blind to user acquisition performance.
Our recent benchmark report found that smaller advertisers with limited resources were forced to minimize their UA channels used in the wake of ATT as performance dropped on some and they struggled to add new ones—whilst the bigger players with a larger portfolio of ad channels and resources diversified further.

On top of ATT’s impact, it’s important to note that at the beginning of 2020 there were 2.7 million apps distributed across Google Play store—the world’s largest app marketplace. Two years since, there are now close to 3.5 million apps in the same place. Despite this massive proliferation of apps, the actual number of apps downloaded by smartphone users is stalling.

This paints a clear picture of app supply increasing and user demand levelling. For up-and-coming apps this means that competition for mobile user attention is increasing, driving up auction prices on ad channels and increasing the Cost Per Acquisition (CPA). This means that the volume of installs you can expect within a given CPA from a single channel reduces. So, increasing install volumes requires the aid of more ad channels than it has traditionally.

Of these challenges facing scaling app advertisers, we have focused a lot on dealing with measurement complexities recently with Appsumer Starter. However, in this post, we want to focus on how scaling apps can diversify channels to reach their desired volume in this era of privacy. Building a bigger portfolio of channels does not only help today it also helps protect install volumes from negative performance impacts when more privacy protocols focused on limiting and redefining user targeting—like Google’s Android Privacy Sandbox—come into effect soon.

How to go about channel diversification

Coming down to it, there are two parts to an ad channel diversification strategy—the when and how. Let us look at how each of these parts work.

When to diversify

When to add new channels is a simple calculation. If you can no longer increase the volume of app installs on your existing channels without exceeding your target CPA or re-evaluating your KPI targets, then quite simply you need to add more channels.

The dynamics of when this is right for an app vary hugely between category, monetization model and geo. However, when we analyze our customer data in aggregate it typically looks like this:

When it comes to diversification, these are some first principle best practices on timing and budget allocation to keep in mind.

Don’t wait

If the ATT slowdown has taught us anything it’s this—don’t wait for performance to plateau before exploring more channels. One of the biggest reasons ATT hit so hard was because app advertisers were less prepared for a world with lesser user targeting. But now that we have an idea of where we’re headed in balancing outcome-based advertising with user privacy, it helps to start early with prospecting new ad channels.

Give each channel time to fire

It’s important to test out each new channel early in the game before it becomes a “need” to diversify, and it becomes too late to media-plan better. This is because each channel takes time to fire and reach an optimal level of performance. Ideally, it helps to have at least 1 channel constantly being evaluated for performance if you’re a scale-hungry app business. Give it a few weeks to assess its real scope of performance for your app.

Set aside 10-20% of your performance budget

Since your next big spurt of growth could come from any channel out there, invest a slice of your performance budget—we recommend about 10-20%—toward testing new channels. To know what’s working for your app (for a particular geo), find the Cost Per Acquisition (CPA) benchmark for your industry/vertical for the channel you pick (ex: Snapchat or Tiktok ads).

After finding the average CPA, multiply it by 2-3x and set it as the ceiling for the first few weeks. The channel slowly optimizes the cost down over time, upon reaching some scale.

How this would look: Let’s say you’re a gaming app and your average CPI is $1 for an ad channel you’ve chosen. Allow a CPI of $2 – $3 for the first few weeks of testing to see how it fires. Ensure you set the weekly or total spend ceiling as your testing budget will allow before you initiate this. Switch off the campaign from the channel only after giving it enough time to show results.

Note: This experiment should be done at a geo-level, as location affects CPI and performance greatly.

How to diversify

While Facebook, Google and ASA are a good place to start, it pays to know what channels can bring you what kind of performance. This insight is pivotal to optimize your campaigns. As self-serve ad channels, Facebook, Google and Apple Search Ads (ASA) can springboard your strategy in the beginning stages giving you an easily accessible and large-scale audience to target.

But as your performance on these platforms plateau, it is natural to look for more niche platforms to grow volumes within a target CPA. Here is how you can evolve your channel diversification based on what stage of growth you are at:

Stage 1: Simplest Scale

This is the period after either launch or an initial product hype in the market. While you may focus on organic strategies like App Store Optimization (ASO) which improves your app page hygiene, we want to focus on paid acquisition which really moves the needle. Your goals at this stage are to reach the biggest relevant audience in the easiest ways. You also want to test the right messaging and creative that you can build off on other channels.

1. Traditional: This approach is a well-trodden path. The philosophy is simple: Look for the channels with the biggest scale, that are easiest to get started with and enable easy testing of messaging and creative. The cadence tends to look something like this, but will vary across iOS and Android (e.g. Google AC may be earlier on Android, while ASA may be earlier on iOS, particularly post-ATT):

Meta Ads: It is a self-serve platform with automated tools to test creative messaging and audience targeting, making it a good starting point. Despite recent performance challenges in the wake of ATT, it is still the most popular channel to start. Even if the scale you can reach may have reduced. Start with broad campaigns using Facebook regular campaigns or Meta Advantage+ Campaigns, continuously narrowing down the targeting to individual interests or lookalikes later down the road, depending on the results. The key with Facebook is that the budget barrier to entry is traditionally low. However, the Privacy Threshold on SKAN for iOS changed that, so some apps are starting to move there later in their growth cycle.
Google App Campaigns (AC): The addition of Google will be earlier on Android and is getting earlier post-ATT. The scale is significant with a self-serve platform and the creative optimization tools are possibly some of the best in native ad tools. It is generally optimum to start with App Install Campaigns to feed Google’s algorithm with the data learnings that it needs, before jumping into in-app action campaigns. A key part is understanding the budget needed to enter. It is generally recommended that for install campaigns you have a budget of 100x your target Cost Per Install (CPI), for in-app action campaigns it is 10x your CPA. Once your campaigns are set up you need to sit back for 2-4 weeks and let Google’s algorithm learn before you start optimizing.
Apple Search Ads: Apple Search Ads (ASA): In the wake of ATT, ASA has likely moved up this list for iOS. The first-party measurement advantages that they have mean advertisers have a better view on performance vs other channels on iOS. That said, like all search channels intent is high but inventory is limited, so costs can be prohibitive, and scale may be smaller for some app types. However, we still see high adoption early on for ASA, likely due to its simplicity and first-party data advantages. A good place to start is running campaigns on exact match to target your brand, competitors and truly relevant keywords (which can also drive-up organic rankings for these keywords). Then run limited broad match discovery to identify additional opportunities that may have been missed.
Verticalized Networks: Often the third channel added might be a verticalized ad network. In gaming it is often Unity or Applovin, for non-gaming it is more likely to be the likes of InMobi or Liftoff. With these options you can find smaller pockets of relevance. However, you will need to have done the challenging work of identifying creatives and targeting with the likes of Facebook and Google before expanding here.

2. Untapped Scale: As a scaling advertiser who wants to move faster, one strategy is to look for up and coming platforms with impressive scale but lower advertiser adoption. Being an early adopter of a big potential channel delivers outsized returns. For example, TikTok had Monthly Active Users (MAUs) in the billions at 1.6 billion and growing fast in Q1 of 2022 along with a high percentage of user’s smartphone time. However, our recent benchmark report shows that TikTok’s advertiser adoption floated around 50%, whilst Google, Facebook and Apple floated around 90%.

So, there was clearly reduced competition matched with significant scale and Matej Lancaric shares a fitting example of how Survivor.io scaled on TikTok in this case study. This strategy can either be complimentary to the traditional approach, or it can be that an app just scales off a nascent channel to begin. Either way, there are not going to be well developed best practices, so you are going to need to firstly know your audience is on that platform and experiment fast with the channel to find the best practices yourself. This strategy is really like betting on the outsider in a race, it is for the brave who have confidence and big ambitions.

Stage 2: Beyond the Walled Gardens

At this stage you need to move beyond what is increasingly the “app advertising triopoly”—Facebook, Google, and ASA. The direction of travel here will vary based on app vertical and audience. However, you have mostly tapped into the large-scale opportunities and have a strong understanding of creative messaging. It is less about choosing channels based on scale now and more about choosing them based on relevance.

Social middle ground

With about 50% advertiser adoption for TikTok and 30% for Snap, these two have very much broken out as the social ad platforms in the mid-tier. Given its fast-growing scale, TikTok is likely to break into the top-tier in the coming couple of years and for some it is an early channel in the mix. However, for now it sits alongside Snap here. If you know your app audience sits on these platforms, they are a logical next step in the diversification process.

The fact that TikTok Ads is projected to reach a fifth of Meta’s ad revenue by the end of 2022 speaks volumes of how quickly it has become a preferred testing ground for ads. With TikTok you can get started with a ~$1k budget to test it out. Obviously, you need to understand the context and refine your creatives for the short form video format of TikTok and if you can tap into trending hashtags you can really increase your reach. Testing TikTok creative with micro influencers and amplifying via paid can be a useful strategy here.

Also, the creative can fatigue in a matter of days, so you need to constantly have new creative concepts tapping into the latest trends. When getting started it is also best to go with broad targeting at a low cost and let the algorithm do the work for you. Then use TikTok’s insights to identify the segments that are performing best.

Snap’s advertiser-adoption post-ATT, has been down as seen in our benchmark reports. However, across geos like the US, UK, France, and India there is scale and in some verticals like dating, the CPAs can be better than core channels. Much like TIkTok, you want to get started with broad targeting here and narrow from there.

Vertical Networks

A niche approach to capturing app users is by approaching verticalized ad networks and DSPs. In the case of these demand platforms, the reach you gain as an advertiser might be more limited as it targets niche inventories but the platforms where the ads show are tailored for higher engagement and conversions. The result is user acquisition numbers that look more conservative but generating higher LTV (Lifetime Value) per user. We often see these platforms come into the growth cycle earlier for gaming apps with the likes of Applovin, IronSource and Unity. Non-gaming tends to be slightly later with the likes of InMobi and Liftoff.

Social long tail

We see advertiser adoption in the low single digit percentages amongst app advertisers for the likes of Twitter, Pinterest, Reddit, Quora, and LinkedIn. This is because the scale is a lot smaller than the bigger players. However, much like vertical networks there are pockets of app verticals where they can deliver lower CPAs or high ROAS. For example, Twitter tends to have an older more professional audience with higher LTVs and higher CPIs to match, that could be right for some apps. Similarly, LinkedIn targeting is strong for high LTV corporate or finance apps.

Pinterest can deliver results for fashion apps and interior design apps who fit the audience and context well. Reddit has niche pockets of scale in the form of Subreddits that will be relevant to different apps and Quora is similar. The point here is these long-tail social networks can deliver high ROAS opportunities for specific types of apps later in their growth cycle.

Affiliates

This is the opposite of a verticalized ad network or DSP which prioritizes quality users over quantity. Affiliates can bring in opacity in knowing where your campaign is scaling and even result in some degree of install rejections. So why do so many large-scale enterprises across the world even pump money into app affiliates? The simple answer is Scale.

Affiliates are connected to an extensive network of publisher ad inventory ranging from direct apps, regional ad networks, DSPs, on-device platforms, app stores, and even other affiliates. This puts an advertiser’s ad across a vast network of supply sources which automatically fetches a more attractive optimal cost-per-action bid. In the initial stages of growth where budget constraints are high, affiliates can offer shot-in-the-arm growth for businesses that depend on app user acquisition for monetizing their business.

Stage 3: Incremental Additions

The biggest challenge at the initial stages of growth is ensuring that any new channels you add are incremental additions to your mix and are not cannibalizing other channels. With 8+ channels in your mix cannibalization becomes a reality. This is where incrementality measurement and testing come into play, to understand the scope of each channel and how it interplays with other ad channels you’ve got in the mix.

Tools like INCRMNTAL and Metric Works are early in their development but help you gage if a channel is adding incremental installs or simply eating up your budgets and hampering other channels from performing. With this understanding you can start experimenting further.

Often this is where those nascent channels with big reach and low advertiser adoption come back into play. A couple of common examples that are cropping up amongst advertisers at this stage are:

Connected TV (CTV): CTV is a prime new example sitting in the space of evolving ad channels. As we stated before, there are now too many apps vying for the attention of users. Given that, app advertisements on Connected TVs help bring products in front of an interest-based audience e.g., Sport fashion and gaming apps for users who consume sports content via streaming. CTVs also help with geographic and time-of-day marketing, to help performance ads reach the right cohort of users at optimal content consumption windows.
Influencers: Naturally, looking for relevant influencers with reach on the social platforms that have performed well earlier in your growth cycle will perform well. However, some apps flip that around and add influencers into their mix early on to test creative concepts and messaging before amplifying via paid media on the same channels. Before you get too bogged down in influencer marketing platforms it is worth testing out the channel with a smaller budget using highly relevant micro influencers. This helps you find the right storylines and ensure that the channel can deliver installs at a reasonable CPI before you move to the mega influencers. Once you nail the influencer channel, starting to scale campaigns with tools like JetFuel help.

Finally: Apples-to-apples comparison is key

As you move outside 1-2 channels you are quickly overwhelmed with multiple ad interfaces, different measurement methodologies and trying to compare performance across channels in spreadsheets.

It becomes harder to quickly get a trustworthy view comparing performance apples-to-apples across channels. Since app advertisers now need to diversify channels earlier, they often do not have the data engineering resources to build a reliable reporting infrastructure to cope with this. Setting up a BI infrastructure even with DIY tools out there consumes a lot of time and resource.

That is exactly why Appsumer Starter exists. To give app advertisers a free and quick-to-set-up solution to compare cross-channel performance on the channels they most commonly use early in their growth cycle. So, before you get started with diversifying channels on your growth journey, be sure to grab your free account here to save yourself time and ensure you have a trustworthy view of cross-channel performance.