MoPub is focused on helping app publishers maximize their average revenue per daily active user (ARPDAU). As many of our existing partners know, we actively provide optimization recommendations to support this objective such as: adding new networks to increase competition, removing low fill network instances to improve latency, and activating Advanced Bidding with our supported and beta networks. MoPub’s unwavering and relentless focus on helping our publishers improve their ARPDAU has created meaningful growth and has also highlighted the need for experimentation.
In 2020, we worked closely with some of our key customers to build MoPub’s split testing framework, designed to open up A/B test opportunities for publishers keen to test the data-driven recommendations that MoPub provides. We are excited to announce that this self-serve tool is now available for all publishers on the MoPub platform.
We designed this feature with three goals in mind:
- Streamlined: We wanted to build something that could be an afterthought for publishers — something that just works out of the box.
- Impactful: It needed to be built to allow publishers to quickly and easily see the performance impact of their experiments, without needing to run complicated reports and invest time to find out that the results may not have been as they hoped.
- Representative: Our publishers require confidence that the experiment results are representative of what to expect when they merge it with the control group.
MoPub’s split test tool allows publishers and developers to redirect requests from a percentage of device IDs for an existing ad unit to the ad unit of their choice. This new ad unit is where they will set up their “experiment”; many of our publishers simply duplicate their existing waterfall and make the changes they want to test.
Although not exhaustive, here are four use cases we observed with our early access customers that can help publishers and developers capture the maximum value possible from this tool:
- Test a new Advanced Bidding network that has become available for alpha and beta (full list here)
- Experiment with different waterfalls for IDFA and non-IDFA users using our targeting feature
- Use different refresh rates for banner ads, as well as different frequency capping for all formats
- Add new traditional networks or custom networks, limiting exposure before expanding to all users
For some of our publisher partners, we’re able to perform overall “APRDAU sprints,” which provide these partners with customized, data-driven recommendations that we feel may help them improve their total ARPDAU. These publishers are then able to test our recommendations by splitting a portion of the traffic to the recommended setup.
Key takeaways to help achieve maximum impact
In addition to the use case highlights above, here are four key functionalities of MoPub’s split test tool, which will help you maximize its impact on your business:
- There is no requirement to make duplicate placements with all of your network partners (although using new placements will give you the ability to validate the results with your external partners).
- Publishers and developers can simply utilize all of the line items and ad units they’ve already created, which will massively cut down their operational overhead.
- No additional SDK integration is necessary, as this feature is server-side. Publishers and developers can experiment immediately with apps that are already live on the MoPub platform.
- With the use of MoPub Analytics, publishers and developers are able to monitor the ARPDAU for the control and experiment groups.
About the author: Marc Bearman, Head of Business Development, EMEA
Marc leads MoPub’s commercial efforts in Europe, Middle-East and Africa. Having spent multiple years on both the Publisher and Advertiser side, Marc has a wealth of experience in the performance space. Outside of MoPub, Marc is passionate about carpentry and sports science.