How 2022 data strategies differ between large, medium, and small app publishers
As we near year-end– amidst the crush of holiday parties, family gatherings, and New Year’s planning– many tend to use this time of year tends for introspection and often find clarity on trends and insights that were lost during the daily grind during a more normal calendar. Looking back on the year, our company’s focus on data has thrust us to the forefront of many data-related challenges faced by app developers during 2021. Apple’s deprecation of the IDFA, the intense market consolidation, a major ad mediator shutting down, and the ongoing need for increasing competitiveness have all fueled the ever-increasing drive for better, more sophisticated data-decision making.
I’ve recently realized these challenges have shaped an interesting trifurcation between app developer types. And while I’ve arranged them here by employee size, it’s not a perfect demarcation for a company’s sophistication and ability to leverage data. The truth is, we’ve worked with 3-person companies who utilize data more effectively than 3000-person organizations. Nevertheless, I’ve arranged the common strategies and aspirations of the varying players in what I hope provides an interesting, introspective view on the strategies and aspirations of how companies plan to utilize data in the new year.
The Large-scale App Publisher
Usually a larger company—over 50 employees— with a mobile app-first focus. A great example are the numerous game companies that are active in acquiring studios, or publishing a massive number of titles. These companies haven’t grown by accident, they are the very companies that are aggressively spending money and resources to actively build their very own Content Fortresses. When we speak to these companies they usually recognize the importance of having a first-party data strategy and have resourced data teams to provide the business with KPIs and metrics to measure and predict growth.
These companies are using and consolidating the massive amounts of data they’ve acquired to drive their competitive edge. They understand that the deprecation of advertiser tracking shifts the ownness for data science onto themselves, they understand and are seeing the need and directional shift for building data clean rooms to consolidate and leverage combined datasets.
As you may expect, the biggest challenge with integrating acquired or partner datasets is the actual construction and maintenance of the data pipeline to achieve the goal of data unification. In a conversation I had recently with a top-10 game company, they put it bluntly “we know the importance of having this data combined, we are just having trouble prioritizing building the actual pipeline.” These companies aren’t generally limited by tight budgets, the strain on resources comes from talent. Another top game publisher told me “Even recruiters are telling me it will take them three months to be able to start looking for my open reqs.” This isn’t an isolated problem: take a look at any large company’s job posting – data engineering is a hot commodity.
Sure consolidation, but to what end? The purpose of aggregating these datasets is for a number of goals:
Decreasing UA cost via cross-promos: This value of compiled users from multiple titles shouldn’t be diminished. If a title is paying $1 per new user and a third of that cost is going to overhead and vendor margin, an intracompany cross-promo strategy could give a 30% price advantage to a UA campaign. This alone could be worth a significant investment.
Creating an asset to increase CPMs: If the above example is how to decrease CPIs for acquiring traffic, the same can be said about increasing CPMs from earned advertising. Privacy changes are making buying traffic difficult for brands but they still need to spend and reach those users. Traditionally, getting the time and attention (read: budgets) of these large brands is reserved for large, trusted sources, running massive campaigns. By consolidating users into audiences at a massive scale, these companies can create their own audiences, and sell those to the advertisers directly. Cutting out the middlemen (Google, Facebook, etc) could increase CPMs easily 30%.
Larger datasets, better models, faster: With large, combined user datasets a machine-learning prediction model can leverage more signals, be calculated earlier, and prove more accurate results.
Collaboration & Shared intelligence: We’ve seen a push for larger companies to bring otherwise disparate teams together to share learnings, what’s working, and what’s not. The year has been seen as a mess of changes – but with those changes comes the need for teams to collaborate on strategies and plans to remain effective.
Usually 10-50 employees, but can vastly differ. On the smaller-end they are usually razor-focused mobile-app developers — on the larger-end headcount side, app revenue is generally an important source of revenue, but not always the primary focus, often seeing internal competing products or business lines. As a whole, both small and large these companies usually understand the fundamental need for data and are in some stage of building out the architecture to support the company goals and aspirations.
These companies recognize the value of data and are in the process of building out infrastructure to support their goals. Generally, they have two priorities: marketing — use data to acquire better, more profitable users; and product – use data to drive user LTVs and retention.
In almost all cases we find the challenges for companies at this stage fall into two camps: application of resources and/or cost. With the number of changes that have hit app developers in the last couple of years, it’s not a surprise that proactive application of data has been prevented from reaching the top of the list. The words of one well-known mid-sized app developer sum it up: “I’d do this today but now I have to migrate off MoPub.”
Second, big data can be expensive – if you’re not careful transforms, processing, and storage can run into the tens of thousands a month. Without a clear ROI, it’s difficult for these companies to allocate the budget to building out the infrastructure.
Leave advanced machine-learning and selling first-party audiences to the upper market: this segment has much more grounded, straightforward goals. Simple business questions that are deceivingly hard – and getting progressively harder – to answer:
Reducing Churn: How can I use data to find where users are leaving my app? Then, how can I prevent or address that problem point?
Increasing user LTVs: What can I do to increase a user’s engagement or conversion to IAP?
What campaigns are driving the most/least ROI? The obfuscation of advertiser IDs has made user-level ROI tough to track. Many will need to figure out models, aggregations or other methods for determining ROAS in 2022.
How can I automate my UA? Many of these developers are clear on CPI goals – achieve 100% ROAS on day 180—they are looking to automate the process of CPI bidding, shutting down or increasing spend on campaigns that achieve these goals.
Small App Developers
These are usually 1-20 person companies. In some cases, they are lifestyle businesses, run by a single of few people. Others have a staff that methodically creates and iterates on the core product, occasionally releasing new apps. In most cases, these companies have achieved a solid niche and are looking for ways to get more sophisticated in how they approach the market.
At this stage in most cases, data is less a strategy and more a means to an end. These companies are looking for a source of truth, the ability to measure user behavior with confidence, and simple answers to hypotheses about their apps and users.
Cost is one of the most important factors for this customer segment, shelling out the budget for tools and vendors isn’t a priority. Very often, what’s offered free is good enough. Unfortunately, the price tag of free often comes with the very limitations that prevent them from answering the questions they need to achieve their goals.
Time and resources at this stage are important too, often the party responsible to implement the technology or uncover the answers is shared and faces higher priority goals that constantly eclipse the ability to be proactive. In the words of a customer: “Finding out 40% of new users never start a game is the single biggest priority for our company. Right after we fix the app from crashing.”
Companies at this stage are looking for core answers and actionable insights from their users. Most often they’re laying foundations to support growth and are building technology and infrastructure to become more sophisticated and help them achieve better understanding and insights of their users.
Source of truth: These companies are usually relying on disparate reporting sources — Firebase for analytics, AppStore and Google Play reports for IAP, ironSource for advertising – most don’t have a clear, unified view of a single user’s action, revenue, or conversions.
Understanding (and influencing) user behavior: I’ve heard questions like “what are users doing before they buy?” Or “how does the onboarding tutorial affect user retention?” Apps in this segment want to change from looking at these events and measurements in silos. Less “15% retention and 30K event counts” and more “Users who opted in are worth 15% more.”
Measuring LTVs: You’d be surprised how many apps still struggle to know the worth of a user over their lifetime. Measuring by user-segment, by user-action? Forget it. Keep in this exercise isn’t the end-goal, it’s a building block to enable the next stage:
Measuring UA Effectiveness: Often these customers are spending on user acquisition with limited ability to track success. They may have a handful of campaigns on a couple of sources – keywords on the AppStore and Google Search – and that’s it. Most don’t know for sure if these campaigns are driving positive ROAS. But all recognize to become effective and start scaling with their growth strategies, they’ll need to start tracking campaign success.
Wherever a company finds itself in the arc of sophistication, the needs and goals for their growth strategy are similar: data is the cornerstone of making the right decision and data remains the malleable glue that supports a company’s growth strategy. While the challenges and costs remain formidable, wringing out signal from your data is the most effective way to make the right decision.
If you think we’ve missed something, I’d love to hear about your experience. Or if you’re wondering how other companies are achieving these aspirations, please reach out or check out some of our other relevant articles.