
So much has been moving under the feet of mobile advertisers in recent time. This has had a significant impact on their day-to-day work and importantly how they generate insights to perfect campaigns. In this post we’ll look at what’s changing and how that’s affecting the analytics and insights stack of a mobile app advertiser.
Part 1: Data normalization: Where business decisions are made
‘Performance’ for an app advertiser can look like different things—Installs, Subscriptions, In-app activities, Purchases, etc.—depending on whether the app monetizes via in-app advertising, in-app purchases, or both. And any app business today will tell you that these performance marketing decisions don’t happen anywhere on the ad serving funnel—not unless a BI tool sits at the very end of it.
That’s right, the big picture decisions aren’t made on measurement tools (MMPs (mobile measurement partners)). They’re not made on media channels either. Optimization decisions are almost always made on BI & reporting tools. This can look like dozens of open spreadsheets, to data analysts building complex BI setups, all in the pursuit of finding the goldmine datapoint or insight that will inform the optimization strategy.
Why is this the case? The simple answer is data complexity. Not only is there now an ocean of data to sift through for performance marketers, but there are also multiple processes involved—sourcing data, pipelining it, cleaning it, aggregating and visualizing, and finally turning it into meaningful insights for all stakeholders in a business.
Here’s how the mobile performance funnel looks for a modern-day app marketer:
Mobile measurement tools or MMPs are key in piping down the opted-in post back data from the publisher back to the advertiser. But they’re seldom built with every single advertiser’s individual complexity in mind. The MMP’s role tends to end with attributing where an install came from, or which ad client is firing more conversions. The data tools that an MMP supplies are not tailored to the specific advertiser’s business goals or delivering relevant optimization insights.
How can they? MMPs are off-the-shelf attribution tools, not custom-built analytics dashboards. BI tools like Appsumer however, are a natural extension to this process as they are built to not just analyze outcomes of performance campaigns but also help advertisers make better optimization decisions. To make performance decisions a modern mobile app advertiser requires the following information:
- Drill-down of campaign performance to understand which optimization levers are working and which ones are not.
- Aggregated metrics from their MMP, analytics data, revenue data, predictive data and cost data from media channels
- Custom visualizations to delve into precise datasets that are meaningful to the campaign and business.
- Full-funnel insights to understand the entire campaign journey from a 60,000 feet decision-making level
BI tools also handle ‘the scale problem’: One of the biggest resources required for an app business to fast-track its growth is time. For a fast-growing app business running campaigns across demographics via multiple channels means exponentially more hours of data crunching. Aggregating datasets from multiple channels and making them all speak a unified language—the process of data normalization—involves copious people hours, which is not readily at the disposal of fast-growing app businesses that are scaling with a lean team.
Part 2: MMP Vs BI: Marrying performance data to the big picture
Let’s get one thing out of the way. MMPs are measurement tools but what they primarily measure and report to an advertiser is campaign post back data—something to tie a successful install to the corresponding ad client who helped deliver it. While this solves for the “which channel gave me x conversions” conundrum, it doesn’t paint the full picture. Half of performance marketing is measurement, but measurement without context is data without meaning.
Let’s understand this with an example. Say an ecommerce app is running a campaign to boost sales by increasing in-app purchases and retarget audiences who are likely to buy the goods they’re browsing on the app.
There are far more variables to consider when putting a number on campaign performance—how the budgets are sliced between channels, what conversion metrics (here it could be add-to-carts and purchases) look like for each channel and measurement partner, and at an even more granular level which keyword, OS, or creative is comparably contributing to these conversion activities.
These performance metrics trickle down from MMPs, SKAdNetwork (SKAN), and Self Attributing Networks (SANs – Google, FB, Snap, etc.) into one aggregated BI dashboard. From here it’s possible to boil down to a concrete ROAS (Return on Ad Spend), to see where optimization decisions need to be made and for this a tool with Business Intelligence capabilities is instrumental.
Here’s how an MMP compares to a BI tool in painting the performance data picture:
Data limitations in an MMP and combatting it
No two MMPs are built the same. While they all solve the same problem, MMPs differ in their approach and interpretation of performance data.
Where this is glaringly clear is in their treatment of iOS ad campaigns. Apple’s SKAN was introduced in 2021 adding an extra layer of complexity into the mobile ad measurement process as it split users into ‘opted-in’ (consenting) and ‘opted-out’ (non-consenting) audiences, creating conversion journeys that can only be partially mapped. This layer of opacity in half the datasets has limited tracking, weakened optimization and user-targeting to a considerable extent.
Since IDFA (ID For Advertisers) helps map conversion events (like in-app purchases or subscriptions) to specific users, the fallout of this was MMPs creating their own Conversion Models. These models were built based on the priorities and needs of app advertisers each MMP works with, and they have their own set of limitations, some more pronounced than the others.
Bridging other limitations
Several MMP solutions are tailored to specific verticals, and hence focus more on the KPIs that matter to businesses in those verticals. Other times it’s because the MMP is not active across all the regions that the advertiser is necessarily trying to expand to. With diversifying of MMP solutions, comes more potential complexity and fragmentation of data across sources.
A BI tool brings in the much-needed order to this chaos, by enabling aggregation, and classification of these disparate datasets. For advertisers, picking up an off the shelf solution that can be tailored to their needs is a definite way forward to manage the data conundrum we’re facing.
Part 3: What’s shifting beneath a mobile marketer’s feet
The online advertising ecosystem has seen a wave of change in the last few years. The first big shake-up came in the form of the pandemic eating away at many businesses and their grow-at-scale plans which meant it was back to the drawing board, as buyer sentiments dipped to an all-time low. Following that, we saw the boom of many online specifically app-based businesses in lieu of in-person shopping, led by an increase in mobile activity across the globe.
If these changes didn’t realign the mobile ads universe completely, what came next sure did—the user privacy revolution helmed by the big two mobile operating systems, Apple and Google. 2021 saw the first of these moves by Apple when it introduced the ATT (App Tracking Transparency) policy for its iOS users and advertisers. This user-focused privacy protocol not only restructured the use of personal data to send targeted advertisements, but also invariably boosted Apple’s skin in the game.
By restricting advertiser access to ad campaign data, creating complexity in performance metrics, propping its own Apple Search Ads as a potential solution to this problem—it upended the entirety of the mobile ad ecosystem, arguably to further its positioning as an ad network and vanguard user data.
Very soon, Google will be following suit introducing its own set of data privacy rules for Android advertisers which will potentially change the way smartphone advertising works forever, given the rest of the non-iOS universe will face a comparable situation where privacy will be consent-controlled by a user, but what privacy means will still be framed by these tech giants.
Where does this leave us?
With more variables like protocols and tech layers between an ad popping up on a publisher app and the mobile advertiser who’s running it. The performance funnel is now more complex and a little broken in some parts. It’s safe to say that measuring the effectiveness of any mobile marketing campaign now rests in the hands of more stakeholders with more stake on campaign data. This makes the advertising data landscape the most complex it has ever been for a modern-day mobile marketer.
User acquisition teams across the board are having to realign their understanding of how acquisition plays out frame-by-frame, in an iron-screen data universe that is soon losing sight of who the target customer is and having to build their own models of understanding based on trial-and-error campaigns, which spells more budget.
But by shifting the data crunching and insight generation work to a BI solution, a lot of the data complexity is simplified while making space on the table for the most crucial factors to assess ad ‘performance’ for a fast-growing app—time, resource, and maintenance.