August 16, 2018
Best Practices for Using Cross-Channel Communications In Your Marketplace
While app usage is skyrocketing, mobile-focused companies still lack best practices for mobile engagement. This is partly due to the real-time nature of mobile: user behavior changes moment-to-moment, and engagement marketing tactics are often rendered obsolete before mobile teams can derive insights and put in place an updated strategy. This is a problem because small adjustments make a big difference when it comes to mobile. With limitless possibilities as to how and when to message mobile users, marketers and product managers need technology that makes automating mobile engagement safe and effective.
Kahuna built RevIQ™ to take the guess work and heavy lifting out of mobile engagement. RevIQ helps mobile apps and brands maximize purchases and conversions by informing them exactly when and how to send messages to their users. Contact us to learn more about RevIQ, or check out the product teardown below.
RevIQ is a new Kahuna product capability that uses collected data and behavioral analysis to automatically optimize push notification campaigns.
There are two components of RevIQ. The first component optimizes send time delay for conversion/trigger campaigns, and the second component optimizes message selection for all recurring campaign types. RevIQ optimizes send time delay by starting with a set of send time delays for a conversion campaign and then measuring the campaign goal achievement rates for those times. The optimizer is updated with the collected data as the campaign is run and progressively adjusts the parameters for the campaign. For example, if a conversion campaign is running and the collected data indicates 75 minutes push delay as the time getting the best goal achievement rate, the optimizer will begin sending more messages with a 75 minute push delay.
RevIQ optimizes message selection based on the campaign goal achievement rate or the user engagement rate if a campaign goal has not been set. The RevIQ optimizer will send a message more frequently if it is outperforming the other messages.
Today, the RevIQ optimizer is available for push notification campaigns. In the near term, we will expand the functionality to optimize email campaigns, in-app messaging, web messaging and additional mobile marketing channels.
RevIQ provides an automated way to optimize push notification campaigns to get the best response without requiring sender intervention. By using RevIQ campaign message selection and send time selection optimization, the sender does not have to manage and manually tune a multitude of experiments. In addition, the automation can respond more quickly and reliably than a human attempting to monitor and update the campaign to maximize the response. As RevIQ evolves it will include more data driven real-time optimizations which would be difficult or impossible for a sender to perform manually.
Further, RevIQ is built to ensure that push notifications are only triggered and sent to users who would not otherwise convert. RevIQ monitors user behavior in real-time and understands the time interval in which 98% of users will convert without intervention. This ensures that the only users who would not convert organically receive a notification, so you aren’t annoying your users with unnecessary push notifications.
RevIQ uses response data collected in a push campaign to update parameters for a Bayesian network that models population parameters associated with the conversion fraction for the send time delay or the campaign message. Initially, all messages and send delay times are considered equal and push messages are sent to each with the same proportion. As the population estimates become better defined, more confidence is gained in the estimates of the best selection of message and send delay time. The proportion of messages sent are adjusted to the combination of message and send delay time which receive the best responses. When the estimate confidences reach a high threshold score, a “winner” is picked and all future messages will use the best parameter selections.
RevIQ requires real-time collection and online learning of parameters to perform effectively. In addition, due to the multitude of modelled and un-modelled variances, adjustments in the learning algorithm are required to make the estimates robust. RevIQ is patent pending, and has been developed and tested over the past few months.
Early results have been very promising. One of our largest e-commerce customers has experienced a 3x increase in purchases using RevIQ to manage the send time delay compared to a long-standing, non-optimized campaign.