So hopefully now you’ve read part 1 of this series on starting a user acquisition (UA) program and set your goals, or at least you’ve got some ideas and are thinking about data to support them. Either way, next you’ll need to consider what analytics you’re going to use to make decisions and judge success.
There are a lot of different solutions that bill themselves as “analytics”, but when we talk about “analytics” in this section we’re really going to be focused on your BI solution (the internal side) and Mobile Measurement Partner, also known as “MMPs” (the external side) .
Examples of MMPs include AppsFlyer, Adjust, Kochava, Branch, Singular, etc., and they more broadly fit into ad attribution analytics. MMPs differ from Product analytics like Mixpanel or Amplitude in that they are more focused on what happens with your advertising campaigns before the install, while still having some relevant after-the-install, in-app analytics to help judge the effectiveness of your ads. They’re critical for successful UA because they are the tool that connects the dots between users seeing and clicking your ads on different channels and matches that to users launching your apps for the first time. While major networks (i.e. Facebook, Google, etc.) offer individual SDKs to measure this, MMPs have the ability to do it across all your channels, arbitrate competing claims between partners, and have tools to help measure owned and earned media impacts.
Given the above, the first question you have to ask in this section is whether you’ll decide to work with an MMP, or integrate individual SDKs. For very small developers, the individual SDK approach can be cost effective and an easier implementation. As UA programs expand though, it starts to become easier to do the work to implement an MMP once rather than an SDK for every new partner you want to test.
The second important question is what you want your system-of-record to be for making UA decisions. This is the place where many developers prefer to use their internal BI solution, while enriching it with data pulled from their MMP. This has the advantage of being more tailored to the specific product or business, while also typically having a high-level of accuracy for things like revenue or engagement. The downside is that there is work required to maintain the integration with your MMP and some translation that has to be done between different systems. Alternatively, a small developer, or a team with a very new product, may decide to rely solely on their MMP to make decisions on UA decisions. This tends to be a simpler and more nimble solution, but is more susceptible to errors due to being completely reliant on how you implement your events and by the nature of using third-party logic.
No two analytics systems ever entirely agree, so you’ll have to determine your source-of-truth at some point. The earlier and more explicitly you do that, the better off you’ll be. Whether you’re deciding whether to double spend on a good weekend, or determine if your product in soft-launch actually has the unit metrics to succeed in a global launch, having faith in your numbers is critical. Continuing with the theme of trust, you’ll be more successful if leadership agrees too, and has bought into your process for allocation (and on what success looks like) Especially when those decisions start becoming expensive.
Now with your goals in mind and your well-thought out approach to analytics you’re ready to start deciding where to put your ads, and what your ads should look like – the topic of part 3 in this series – coming soon.