Last week I wrote about why ad creative has become such an important part of the mobile marketing workflow and provided a general framework for producing performant mobile ad creative at scale (the content is an adaptation of what will be presented in the forthcoming Modern Mobile Marketing at Scale workshop series taking place in October). In that article, I provided some background on the tectonic shift that has taken place within mobile marketing over the past two years as ad platforms have prioritized algorithmic campaign management, and while the piece explores the role that ad creative plays in rapid experimentation of both audiences and the ad creatives that best resonate with them, it doesn’t explain the role that experimentation itself plays in this modern mobile growth paradigm.
The process of experimentation generally appears to be siloed within different specializations at many companies: with ad creative, or with event testing, or with in-app content personalization, etc. In other words, while experimentation is at the very heart of what growth teams do, many teams see it not as a functional process in and of itself but rather as an abstract exercise that they apply to specific aspects of growth. For many developers, this results in various components of “growth” being broken out into disparate teams that operate independently across the user lifecycle:
This arrangement — a functional separation between acquisition and in-product “growth” teams — is not efficient, and it will become extinct as more and more companies adapt to the operating standards of the new mobile distribution environment. But before exploring that, two notes about the above diagram:
- The acquisition team pyramid as depicted here represents how many advertisers today may think about team composition (with the size of each layer corresponding to the amount of resources / headcount allocated to it), but that is rapidly changing, too. Media buying as a distinct role is becoming almost obsolete as analysis and creative experimentation become the high-impact levers available to marketing teams. The acquisition team of the future looks more like the below, with data science and analytics teams driving marketing spend and with media buying being potentially completely automated away:
- The impact of A/B testing-oriented “growth” teams is, I believe, generally exaggerated. My belief is that endless A/B testing, when it is constrained entirely within the product and not connected to marketing, can actually destroy value as it pushes users into local maxima and ignores opportunities for personalization. I have seen growth teams thrash on product iterations so aggressively that experiments confounded each other and almost certainly subvert the work of acquisition teams in terms of finding high-potential users.
Those points out of the way, it’s useful to consider how the siloed teams depicted in the diagram amalgamate and evolve collectively into the growth team of the future. In order to do so, it’s important to identify what I believe to be are the critical modern functional components of growth:
- Media mix modeling. The increasing opacity of placements on Facebook and Google as well as the general growth of non-direct response budgets for mobile advertisers (which is in part a reaction to the aforementioned opacity) is creating a situation where per-user marketing attribution is impossible or impractical. As mobile moves into a post-attribution era, media mix modeling is becoming the means by which mobile advertisers make budgeting decisions. Media mix modeling is mostly a data science exercise that informs media buying, but it also has a tremendous impact on product metrics. Media mix modeling and marketing incrementality are the at the frontier of exciting marketing science at the moment, and these functional skillsets will sit at the core of the growth team of the future;
- Content personalization. A/B testing is a blunt instrument and, as noted earlier, tends to not work well or even be counterproductive when it isn’t aligned with what an acquisition team is doing. Rather than optimizing the entire user base as a monolithic cohort, I am seeing the most sophisticated advertisers now personalize the user experience on the basis of many demographic features that include the user’s acquisition channel. With this approach, signal from the source of acquisition flows through the app install and results in a personalized experience for everyone based on data that is gathered throughout the entire user lifecycle (and not just the portion that is spent in-app, after acquisition);
- Event signal measurement. As algorithmic campaign management obfuscates audience development and targeting, advertisers increasingly rely on in-app events sent back to campaigns in campaign strategies like Google UAC and Facebook AEO and VO to help ad platforms identify high-value users and target them disproportionately. Part of scaling advertising campaigns is experimenting with these events to find the ones that deliver the best, most meaningful signal to the ad platforms, which requires tight coordination between the user acquisition and product teams;
- Creative experimentation. As the second set of fundamental inputs (along with in-app events) of modern mobile advertising campaigns, creative experimentation is a critical aspect of scaling mobile growth. I cover this in great detail in this post.
These two practices are obviously inter-related: as media mix modeling informs budget allocations across channels and media formats based on overall, top-line revenue, content personalization optimizes the experience at the user level on the basis of information that is generated in the acquisition process (eg. users from acquisition channel X are initially placed on content path Y within the FTUE). In an article titled A/B testing can kill product growth, I visualized the approach that most growth teams take in A/B testing product features and user flows:
With content personalization that includes acquisition information, users are put on content paths on the basis of their source of acquisition from the time of install. But what’s more, the A/B testing mechanic, which is very manual and prone to superfluous specificity, is replaced by online bandits mechanics in most cases of modern personalization, which are easier to manage and scale. Implementing this kind of personalization is clearly not so much a product or marketing exercise but rather a data science activity: building, measuring, and tuning these models is at the heart of the process. And media mix modeling is no different: a model tests budget allocations for incrementality and overall return and adjusts it accordingly.
So what does this mean for the growth team of the future? It means three things:
- The team has oversight of both acquisition and in-app behavioral optimization;
- The team is focused on building systems that determine optimal content paths for each user;
- The team recognizes that it has opportunities to drive improvements in product LTV not just by optimizing acquisition but also by optimizing the product experience at the user level and finding the best in-app signals of value to transmit to ad platforms.
In practice, this means that the growth team — as a singular, integrated entity — has broad oversight over both acquisition and in-app content optimization that benefits from visibility into acquisition source:
So just as ad creative has grown in importance to necessitate a fundamental change in the way growth teams operate, so has experimentation: it has extended into the app, requiring a growth team’s purview to be enlarged past the “acquisition event horizon,” or the point after which a user downloads an app.
One question that the above analysis raises is: why is this the future of mobile growth and not the present, if these systems are already in place? One reason is that, generally, teams are slow to adapt to significant ecosystem changes and they do so very gradually. A mobile-first company just being formed right now might build its growth team in a way that facilitates this modern reality, but existing teams must re-tool and re-hire.
A second reason is that there is a limited talent pool of growth marketers with the skills needed to excel in this environment, and they are in great demand. Hiring a competent growth marketer with a background in scaling mobile products into millions of dollars of advertising spend per month (or millions of DAU) is difficult; advertisers struggle to build growth teams and often run into the chicken-and-egg problem of mobile advertising. But this new paradigm undoubtably is the future of mobile growth, and whether slowly or wholesale, teams must prepare for and adapt to it.