It’s no secret that the FeedMob team thinks SKAdNetwork (SKAN) experimentation is key to a well-rounded mobile growth strategy outside of search and social media, but measurement for down funnel events still poses several challenges. However, we’ve gained several insights by working with our clients to test SKAN and are ready to share them.

Does SKAN work with all partner types?

Simply put, no SKAN is not compatible with all of our partner types. There is one main barrier many partners are unable to overcome. In order to be compatible with SKAN, the partner needs to touch the supply (run SKAN campaigns directly through their SKAN ID, which allows for the SKAN postback from Apple to be received). For this reason, rebrokering and affiliates are not SKAN-friendly, but partners whose ID is implemented (like DSPs) are able to receive postbacks. 

As FeedMob’s CEO said in a previous Q&A, “As we move forward to a SKAN-based attribution model on iOS, it is clear that as a tech-driven agency, we need to adapt. Attribution information now flows directly to our supply partners; with that in mind, we’re building reporting and analytics based on MMP APIs as opposed to purely postbacks to our endpoint. I believe this will allow us to tap into supply that previously was not available to our clients and create more sustainable spend and analytics aggregation across the iOS landscape long term.”

Is SKAN impacting pricing?

Many app marketers wonder if — and how — SKAN impacts pricing. The short answer is, “No.” Of course, there is more to the story than that.

• If we’re running a flat CPI campaign, our clients are typically not willing to pay more for SKAN traffic right now.
• If we’re running a CPM/CPC campaign, we could see the cost of user acquisition significantly change between the two.

One may also be tempted to assume that SKAN traffic is cheap, as it can be difficult to attribute — but that does not seem to be true either. In fact, we mostly see the opposite. Because of SKAN’s postback logistic restrictions — which means some users who convert are missed and not attributed, making CPA calculations a bit misleading — it could actually be more expensive in terms of CPA vs. non-SKAN campaigns (where we get posted back on events in real-time, and there isn’t such a strict time frame where the next post-install event [PIE] needs to occur). However, even with this, we are still paying out at a comparable rate for SKAN users. 

With that in mind, if a client is not willing to run probabilistic VTA, CPM user acquisition on SKAN might be even more cost effective since that’s a group of users the MMP probabilistic campaign can’t attribute. Ultimately, outside of flat CPI campaigns, there are so many variables that pricing really depends on what type of campaign you’re running (which could potentially perform better or worse on SKAN).

When does SKAN work best?

The nuance to note when it comes to SKAN is that tracking down funnel events is difficult. In other words, SKAN is suited for tracking installs and other early-conversion events. After the first PIE, users must take another action within 24 hours to keep the clock ticking. This means attributing events that happen later in the user journey can be difficult to track. If the conversion on the following PIE occurs after 24 hours, then we will not receive a postback of the second PIE. With this strict time limit, we likely are missing out on further down funnel events.

That’s why adjusting your campaign goals to SKAN’s strong points is important. You can talk to one of our account managers to learn more about how to structure your campaign goals for SKAN success. Alternatively, if you want to learn more about the general options available to you for growing your app outside of search and social media — and how FeedMob attributes those campaigns — reach out to set up a free consulting session.