Tracking iOS User Acquisition ROI: To SKAN or not to SKAN?

This 2023 article explores the measurement questions around iOS 17 and SKAN 4.0, as those changes were being discussed. For a current UA decision, separate recorded product revenue, campaign-attributed conversions and modeled estimates. The useful question is which evidence can support the budget decision, with its coverage and limitations stated clearly.

The Crossroads of iOS User Acquisition

Tracking ROI used to be a no-brainer, but now it’s a whole different ball game. Marketers are left scratching their heads: Should we go the easy route with SKAN, deep-dive into the maze of opt-ins and data modeling, or try juggling both? Here are three potential scenarios to help marketers determine the best path forward.

1. Going All-In on SKAN

SKAN provides privacy-limited campaign attribution signals. When comparing them with MMP reports, document the events, reporting cutoffs and rules used to handle overlaps. Ask the vendor which conversions are deduplicated and which values are modeled. A combined report is useful only when its definitions and limitations are clear enough to support the intended decision.

2. Juggling Opt-Ins and SKAN Data Modeling

The second scenario is to focus on obtaining opt-ins from users while modeling the rest of the data. This strategy gives you access to detailed user-level data for those who opt-in, allowing for more precise targeting and measurement. For instance, using MMP attribution for post-install metrics & revenue and extrapolating this data with SKAN installs can provide cohort ROAS data and CPA, both at a global level and per channel. However, this approach presents challenges. It is hard to implement and there’s no guaranteed way to check if extrapolations are accurate. Moreover, under this scenario, marketers are subject to changes in opt-in rates and user behavior, which can significantly impact the amount of data available for modeling.

3. A Hybrid Approach

The third scenario involves a hybrid approach, using a combination of SKAN for campaign attribution and opt-ins and data modeling for deeper insights. Companies like INCRMNTAL and MetricWorks that focus on incrementality can help assess channel performance, identify the additional metrics it brings on top of the existing ones, and spot any cannibalization. Although this approach offers the best of both worlds, it also requires careful implementation and ongoing management. The models need time to be trained, and there’s no easy way to reconcile the incrementality approach with the classic MMP attribution.

iOS User Acquisition ROI Tracking

Not just a bunch of letters, but a really complex task. Let’s explore some comprehensive methods to help you navigate this terrain and track your iOS user acquisition performance.

1. Looking at Overall Marketing ROAS

One essential tactic to consider is to evaluate your overall marketing ROAS. This involves a simple calculation of your total app revenue divided by your total marketing spend. Comparing this figure against other operating systems or against listed companies will provide valuable insight into your positioning within the industry. Notably, this indicator is unaffected by App Tracking Transparency (ATT). Yeah, you heard that right. Your spending? It assures you it’s on point. But there’s more. It doesn’t just work in a vacuum. You can compare it across different operating systems. You can even size it up against what the industry bigwigs are doing. Sounds like a silver bullet, doesn’t it? Hold on, though. It’s not your magic wand for budget allocation. It’s more like your sanity check. In the realm of User Acquisition (UA), cohorted data is king. This KPI? Not so much. And here’s the kicker. If the quality of your installs takes a nosedive, this KPI won’t flash red lights. It’s not your early warning system. So while it has its perks, it’s not the be-all and end-all. Keep your eyes wide open.

2. Looking at Overall eCPI

Another key strategy involves examining your overall eCPI (Effective Cost Per Install). This requires analyzing your marketing spend over total installs, attributed or not, to see if your marketing efforts are driving more or less installs over time. This method has its pros and cons. On one hand, it’s unaffected by ATT and easy to observe over time. It factors in the organic impact of your marketing spend, thereby adding the organic installs. On the other hand, it does not inform you about the profitability of installs and cannot help allocate budgets between different channels.

3. Combining SKAN + MMP Data

A Mobile Measurement Partner report may combine several attribution sources. Confirm how it handles overlaps, missing signals and modeled values before adding totals or comparing channels. Use consistent event definitions and cohort ages, and keep the recorded and estimated portions visible. Choose supported measurement methods that respect the applicable platform requirements.

4. Extrapolating Data

Extrapolation estimates outcomes for data that cannot be observed directly. Record the sample, assumptions and uncertainty: users whose data is available may behave differently from those whose data is missing. Compare estimates with product-level revenue and mature cohort evidence before using them for allocation. A model is a decision aid, not a replacement for its missing observations.

5. Embracing Incrementality

Lastly, an increasingly popular method among companies like INCRMNTAL and MetricWorks is looking at incrementality to assess channel performance. This approach allows you to identify the additional return a channel brings on top of the existing ones and detects cannibalization, which can lead to negative impacts. Despite its benefits, this method does require a considerable amount of time to implement and train the models. Also, reconciling the incrementality approach with the classic MMP attribution is challenging. Some tools even ask you to temporarily pause channels x geo to enrich the learnings and increase accuracy.

Conclusion

The era of iOS 17 and SKAN 4.0 is here. Tracking user acquisition on iOS has gotten more complex. But don’t panic. First, let’s talk ROI. Knowing your Return on Investment isn’t just a good idea anymore. It’s essential. You need to weigh the pros and cons of each marketing strategy. But it’s not just about numbers. It’s about being nimble. The landscape has shifted, and flexibility is your best ally. Be willing to experiment. Try out different approaches to see what sticks. Above all, remember who’s at the center of all this—your user. The human element shouldn’t get lost in the mix of metrics and strategies. Keep the user journey at the heart of your decision-making. And in this ever-changing environment, that might be the most stable strategy of all. As our Performance Marketing Director points out:

 

“There is no single way of measuring iOS UA paid actions. In this fast-evolving time, we need to embrace all the data we have, understand its limitations and define our own approaches until we find something we are happy with. It doesn’t matter if it is SKAN only, data modeling, a mix of both or any other solution.”

Pablo Perez | Performance Marketing Director @ Admiral Media

 

The decision to go all-in on SKAN, pursue opt-ins and data modelling, or adopt a hybrid approach depends largely on the specific needs and resources of each app. While SKAN provides a simpler and more privacy-compliant solution, opt-ins and data modelling offer more detailed and earlier insights at the cost of greater complexity. The hybrid approach could potentially provide a balanced solution, but it requires careful management. The key is to understand the pros and cons of each scenario, and align the strategy with your overall business objectives and capabilities. When it comes to SKAN this is just the beginning of the road. If you are just starting this journey, we’d love to help you boost your performance marketing efforts for your app. Get in touch with us right here.