mParticle, SDK connectors, and advertising


An increasingly important company in the mobile app ecosystem is mParticle. This company was founded by Michael Katz, who founded Interclick – an early-ish user-data-focused ad network that Yahoo! bought for $270M in 2011. mParticle was founded in 2013 and has since raised $72M. Their chief rival is a company called Segment.


The main idea behind mParticle (and Segment, largely) is that apps have many integrations like analytics, attribution, push notifications, etc. and each has their own SDK. Integrating with a new provider, learning if it’s a fit for the app, and evaluating the ROI are all expensive propositions. mParticle removes the need to install separate SDKs to work with various vendors, and instead creates connectors to the various services, and also facilitates data warehousing for the stream of data that the app generates through connectors to the warehouses themselves.

An example is the connector for AppLift. AppLift is a mobile app re-engagement platform (AppLift shows you ads if you’ve installed a mobile app but haven’t used it for some time, they try to show you ads in other apps to get you to use it again). If you want to use AppLift, you need a business relationship with them no matter what. But then you could either install their SDK, meaning try to get engineering time at some point in the future, or you simply enable them in the mParticle app. You would add your AppLift “key” in the mParticle UI, then the mParticle SDK (which you have already installed) would automatically forward all events server-side to AppLift that would enable that service. This means that mParticle is receiving a stream of all the relevant events and forward the appropriate ones on to the appropriate services. If you were A/B testing between AppLift and a competitor, the engineering resources would be even more complex - but this is out-of-the-box for mParticle. 

The service itself isn’t particularly complicated – but the value is in the sheer number of integrations and the turn-key implementation of those integrations across a vast number of platforms (iOS, Android, Roku, Amp, Unity, React Native, Xamarin, etc), and the fact that relatively complex rules can be setup that unify different platforms through different outputs and audiences. 

mParticle does not show ads itself – its primarily functionality is on the data side. That said, one could imagine a company that does something similar in the app context for ads. If you create a relatively standardized set of advertising components (banner ads, video ads, native ads, interstitials, etc) and a fixed set of what data points every advertising SDK requires (pixels on render, viewability, video start, complete, billing events, etc), then a single unifying SDK could be used to automate integrations with dozens of app “SDK” partners. This would make it easier for companies like TripleLift to gain market share. It’s unclear why this doesn’t exist yet, though it’s likely very complicated and ad companies likely encrypt at least some portion of the response that would make this more complicated. That said, there’s at least one company -  that seems to be doing something similar.