Build or Buy: How independent apps can scale ad performance in an era of data complexity

The mobile ad ecosystem has been shifting under our feet over the past year. If you run mobile performance ads you’ll be aware that the advertiser playbook is being ripped apart. The industry is experiencing a wave of change right from consent-based targeting, shifting ad budgets, and how businesses look at ad data—with the latter leaving a significant impact on performance advertising.

A few tech giants are at the forefront of these changes. Apple dropped its iOS privacy framework ATT (App Tracking Transparency) in 2021, effectively galvanizing the entire mobile advertiser ecosystem toward a future without user-level targeting. Google for its part recently announced that it would roll out its own mobile privacy framework—‘Privacy Sandbox’ for Android. This will take a crack at limiting the Android user identifier GAID limiting advertiser control over consumer data. Much like Apple’s consent-driven landscape, if potentially a little more advertiser friendly.

For advertisers this translates into more focus on unifying fragmented data that’s now coming in from multiple sources:

a. Cost data from media channels

Self Attributing Networks (SANs) like Facebook, Snapchat, TikTok
More traditional ad networks and DSPs like Applovin, InMobi, Unity Ads
Add into that many newer channels like influencer marketing and Connected TV (CTV)

b. ID based attribution data from mobile measurement partners (MMPs)

c. Non-ID based attribution data on iOS from SKAdNetwork (SKAN) and shortly Google’s Attribution Reporting on Android

d. In-app analytics data from the likes of Amplitude and Mixpanel

e. Increasingly advertisers are also turning to modelled data to fill in the blanks on measurement data losses to privacy moves with probabilistic attribution, Media Mix Modelling (MMM), or incrementality measurement.

Performance ads that usually cover down-funnel business activities like user acquisition, app monetization, and in-app spending are severely hit with the weight of these changes. The first reason is that performance ads largely depended on user targeting to ensure each ad dollar is spent on generating a business outcome from the exact target audience. Two is that each stakeholder in the mobile ad space now has their own set of data to show how these ads are performing, none of which is an apples-to-apples comparison across channels and OS’s.

What this means is data is now more complex than ever. Naturally, assessing an app campaign’s performance has more steps now. Making sense of this complex data as a small business results in multiple dozens of open tabs and spreadsheets, hiring or outsourcing data experts, culminating in figuring out what solution they need to solve for this time drain and loss of visibility.

So how does this affect independent app companies?
You know, the businesses which can’t double or triple their budgets overnight to hire a small army of data engineers and scientists to solve it.

Impact on Independent App developers

A deeper look at the performance numbers since iOS 14.5 ushered in ATT—removing user identity and introducing a degree of data complexity—reveals a markedly higher number of smaller scale advertisers have suffered at the hands of this policy change compared to enterprise businesses. A study conducted by Appsumer found that advertisers with ad budgets in the millions are finding more success on iOS post-ATT whilst those with lesser budgets especially under $250,000 a month have been impacted significantly more. Here are some of the challenges they face.

a. Mobile ad data is more chaotic than ever before: A large part of the scaling problem is making sense of performance to find opportunities.. With a bigger set of resources, the largest advertisers have been able to quickly build data infrastructure around Apple’s SKAdNetwork (SKAN) attribution solution to get a better understanding of campaign performance.

But not all independent app companies have a small army of data engineers and scientists to crunch performance data to optimize for scale quickly.

b. Privacy thresholds hurt smaller scale advertisers most: A key part of SKAdNetwork is the privacy threshold, which is a mechanism to anonymize SKAN postbacks. On Facebook, unless your campaign has greater than 128 installs per day the SKAN postback will include no information on conversion events. This threshold is lower on other channels, around 20-30 installs per day based on what we can see, due to Facebook’s SKAN mechanics. However, it is a big challenge for smaller advertisers to reach the scale required with their budgets.

c. Less diverse channel mix: Smaller-scale advertisers have had less of a need to diversify their channel mix up until now running on 2-3 channels with their scale of spend. However, with sweeping privacy changes some channels have been hit harder than others. For many smaller advertisers if the performance for one of their channels was hit hard their spend reduced. Larger advertisers have a more diverse channel mix meaning they saw less macro impact.

Solving the data challenge

At the heart of the problem sits data. A robust data infrastructure to make sense of this data fragmentation is a necessity at this point. With partial data, structured differently across multiple sources a new data layer is required to manipulate and model data and present this performance view in a digestible format to make effective performance decisions.

An effective BI solution can bridge the gap between fragmented and incomplete data by normalizing and unifying it for apples-to-apples comparisons across channels and OS’s in dashboards and reports that can be interrogated by User Acquisition Managers.

If you’re an independent app business, this decision of whether to build or buy this data layer is a lot harder as you have pressure to scale fast and don’t have the small army of data scientists and engineers that larger scale app businesses do. We’ve put together a table that can help you navigate this decision with more ease.

It’s evident that the answer to buying or building boils down to one big variable, you as an advertiser. And while building an in-house solution from scratch can help streamline performance marketing data, there’s a bigger cost to pay in terms of the time spent in addressing feasibility over prioritizing growth.

Buying an off-the-shelf mobile ads BI solution like Appsumer can help growing advertisers not only get the right optimization insights but also can help them redirect the building and maintenance manhours to solving the bigger challenge at hand—scaling.