The 4 Axes of Accurate Retargeting

By Director of Data Science, Ido Zehori

Bigabid has built an A/B testing system that is able to run hundreds of experiments simultaneously to provide answers about the effect size and statistical significance of creatives and treatments to any sub-group.

The 4 Axes of Accurate Retargeting

By Director of Data Science, Ido Zehori

Intro

Bigabid took on the challenge of retargeting only a couple of years ago. At that time, it wasn’t nearly as widespread as it is today. This allowed us to take some novel approaches that are paying off today. Together, with our design partners, we’ve tackled many issues that arose mostly around these four main axes:

  1. Establish the most valuable user segments to target and when to start targeting them
  2. Measure the activity (incremental experiments, different KPIs for different segments, etc.)
  3. Create and test creatives with custom messages and special offers for the various segments 
  4. Technical Integrations (MMPs, audience tools, deep linking, etc.)

 

In this article, we will tackle them one by one and uncover the solutions Bigabid offers.

1. Establish the most valuable user segments to target and when to start targeting them

First, we research and analyze the client CRM to accurately determine the NNR (natural return rate) or different user segments with respect to idle days (days with no app open), or various other segmentation methodologies that are relevant for the campaign’s goal and the nature of the app. Then, we divide the users in the app into different segments and offer a reach analysis for each group alongside suggested KPIs for the group based on data from the stream in the real world.

For more on user segmentation in retargeting check out The 4 Pillars for Successful Retargeting.

We’ve built a powerful data integration pipeline that is able to process the events coming in from the app in near real-time. This engine provides us with an unmatched capability to segment your users according to hundreds of different features in order to serve our algorithm data that is constantly updated and refreshed for maximum results.

Discover more about Real-Time Data Analysis.

2. Measure the activity (incremental experiments, different KPIs for different segments, etc.)

Bigabid offers periodical meetings with our teams to provide reports and valuable insights from the campaigns once they are live. Such meetings are mutually beneficial. Then, we provide an incremental impact measurement framework with the experiment design and analysis. 

Learn more about how Bigabid measures incrementality in Incremental Value.

3. Create and test creatives with custom messages and special offers for the various segments 

Bigabid has built an A/B testing system that is able to run hundreds of experiments simultaneously to provide answers about the effect size and statistical significance of creatives and treatments to any sub-group.

Discover more on Bigabid’s approach to retargeting ads in 5 Tips for Successful Retargeting.

4. Technical Integrations (MMP, audience tools, deep linking, etc.)

Bigabid’s CS team has accumulated experience with numerous clients that have been successfully onboarded across all major MMPs. We all work together through the process and make sure you won’t repeat any of the mistakes we made during development and innovation.

Learn how Bigabid partners with MMPs here.

These important tools are an addition to our already super powerful ML (machine learning) architecture that is in charge of 100% of our media buying. Some key features in our algorithm are:

  • Features about who the user is. Things like-  
    • What interests them
    • How passionate they are about different types of games
    • Is the user ad-blind, and to which ad types do they respond best etc.
  • Features about the environment the ad is presented in-
    • Is the ad clickable or even viewable?
    • Is the ad visually appealing?
    • Can the user install the app in their current environment, etc.
  • Features around natural return rate; we only want to target users that have a low probability of returning naturally to the app. We have a model predicting what the chances are of the user returning to the app without treatment. Then, we optimize for the difference between the chances to return to the app with and without showing the user an ad. By doing so, we optimize the incremental value of the campaign.

Conclusion

Bigabid’s retargeting practices have been refined over a relatively short time though we’ve had more time than most of our competitors as we developed the process before it became more widespread. Having experienced how fast things have evolved, we look forward to continuing to spearhead retargeting technologies with drive and unshakable transparency. 

Read about Bigabid’s core values in Partnering with Transparency. 

Transparency not only in sharing insights and refining with our clients and partners but with our potential future clients and the mobile advertising industry as a whole, hence articles like this. It’s our confidence in our technologies that allows us to be so transparent. Come be part of our retargeting success.  

Results based advertising starts here.

Contact us
Please fill out the form below to submit your interest.
Please review our privacy practices: read privacy policy.