Predictive Bidding — you can now let your data tell you which users are most likely to convert

Finding ways to engage mobile app users remains a top challenge for mobile app performance advertisers and engaging with users that already have the app is a great way of driving retention.

The problem is that when creating different user segments to target for engagement campaigns, marketers have mostly relied on intuition… “users that haven’t made a purchase in the last 30 days”, “iPhone 6s users that open the app frequently but don’t purchase, etc.”. This sometimes works, but often people within those groups have different characteristics that make them more or less likely to end up making a purchase.

That’s why we are pleased to announce Jampp Predictive Bidding, a powerful automated new way to leverage your users’ data to predict which of them are more likely to convert and then device custom bidding and creative strategies to engage them.

If they’re likely to buy, we’ll find them

With Predictive Bidding, we use predictive models and real time data analysis to determine the likelihood of conversion for each of your non-activated users. The more data and conversions we analyze, the more the platform learns about what makes a user convert.

To do this, we start with a global integration of your user data so we can understand who your top customers are and what they have in common from a behavioral and contextual point of view. We match the profile of your ideal customer to similar users to determine how likely they are to convert, considering a number of behavioral and contextual variables such as previous in-app behavior, demographics, device information, and more.

Each user that hasn’t converted is grouped into one of the following predictive segments:

• Very likely to convert
• Likely to convert
• Less likely to convert
• Unlikely to convert

In the dashboard, you’ll see how big each Predictive Segment is and its associated probability of conversion.

Finally, our platform suggests variable CPCs for each segment, which are updated in real time. This focuses your budget on the best users, driving more conversions and ultimately lowering your CPA. Within our traditional engagement plan, our full funnel analysis has a number of stages, as shown in the below example.

With Predictive Bidding, the bids are variable based on how likely a user will convert. For example, in the diagram below, we are showing the newly acquired group of users will be split into four predictive segments with variable bids of $0.25 — $0.50. This assures that you pay higher bids for those users we know are more likely to convert and less for those who are not as likely.

The segments are updated automatically every day as new user data comes in.

Why Predictive Bidding

Early results from our beta customers (here is a case study with a large e-commerce app) are not only showing higher volumes within a 7 day period, but their CPAs have decreased by over 30%.

Here’s a look at some of the numbers:

Whether your customers buy tickets, book hotel rooms, listen to music, book car rides, or anything in between, our platform can help you predict the users who are more likely to convert, increase conversions, and lower your CPA.