Now, let’s unravel an issue that has confused many app marketers. While IS leverages probabilistic models for optimizing advertising campaigns, a rift arises when it comes to billing, where SKAN data steps into the spotlight. The result? Inflated expenses and misconstrued campaign performance. Let’s dissect the intricacies and unveil the recommended course of action.
Understanding the Data Dilemma
IS employs probabilistic models to optimize campaign performance. However, here’s where the challenge emerges – when it’s time to settle the bill, SKAN (SKAdNetwork) data takes center stage. The repercussions are not negligible. If you rely solely on probabilistic models for budgeting, performance analysis, and future planning, you might find your end-of-month invoices higher by 20-30% more than you initially budgeted for.
Here is a real example:
This gap could have a substantial financial impact, especially if your finance and user acquisition teams operate in separate silos.
The reasons behind this discrepancy are rooted in the inherent differences between SKAN and probabilistic attribution models. The platforms do provide documentation on their billing and reporting mechanisms, but this crucial detail can often be overlooked, particularly for those new to mobile advertising or utilizing self-serve platforms.
Finding Solutions: Two Approaches
So, what should you do to navigate this data dilemma effectively?
- Aggressive Approach: Stick to billable spend, but apply a multiplier (roughly 30%) to your reported conversions. This approach assumes that probabilistic models might miss out on approximately 30% of actual conversions. While this approach potentially overestimates conversions, it aligns more closely with your actual spend.
- Conservative Approach: Use SKAN conversions as the definitive source of truth (instead of probabilistic models) and align your billing with this data. This method relies on SKAN conversions and billable spend as the most objective and dependable sources. It prioritizes data accuracy over inflated conversions.
| Spend | Conversions | CPA | ||
| Aggressive | $73,872 | 2,914 | $25.35 | <- take billable spend, but add ~31% to your Conversions, based on the assumption that probabilistic is actually missing out on 30% of conversions. |
| Conservative | $73,872 | 2,211 | $33.40 | <- Take billable spend, but add ~31% to your Conversions, based on the assumption that probabilistic is actually missing out on 30% of conversions. |
However, a potential challenge arises when you discover the absence of SKAN conversions directly within the ironSource (IS) platform.
Considering the complexities involved, we recommend adopting the conservative approach. Embracing SKAN conversions and billable spending as your reference points ensures utmost data accuracy, allowing you to make informed decisions and facilitating fair comparisons across different advertising channels.
