Anatomy of the Adoption Phase

For the most part, the adoption phase consists of the following steps:

Fill out the Adoption Plan (provided by your Customer Success Manager)
Finalize initial experiments to ensure data is reasonably calibrated
Ensure data will be imported into Polaris at your normal optimization frequency
Ensure Polaris's incrementality metrics are available to all people and tools necessary to execute the normal marketing processes
Begin executing the normal marketing processes (e.g., bid and budget optimizations) across all channels in countries participating in the adoption phase using incrementality data while continuing to use last touch data in non-participating countries
Analyze the performance impact after the adoption phase is complete
Decide on the future use of incrementality measurement and experimentation

Adoption Plan

The Adoption Plan helps you set the scope, success criteria, and expectations for the adoption phase. Here are the most important questions that need to be clearly answered in the Adoption Plan:

Which country or countries will you select for participation in the adoption phase? Recommendation: Select countries with larger budgets so performance impact is more detectable and keep in mind that how campaigns are currently configured to target groups of countries can play a major role in which countries can be optimized separately from others.
Which early KPIs are most important to you? Examples:  D3 ROAS, D7 ROAS, Installs, D1 Revenue, etc.
What is your objective during the adoption phase for evaluating Polaris and its value? Examples: Maintain or improve country-level D3 ROAS or obtain ground truth incrementality via lift experiments, which are no longer possible on iOS, etc.
Other than marketing, how else do you use last touch data currently? Does last touch data feed into any CRM tools, financial reports, LTV prediction models, BI systems, etc.? Ideally, incrementality data replaces last touch for those applications as well, but it isn't always possible to cover everything in the adoption phase.
How will you import data into Polaris with the same frequency that you optimize? If you are using a data partner integration, data imports are automated, but if you are importing data via custom CSV, you'll need to ensure imports are happening regularly either manually or via script.
Does your existing marketing process handle significant swings in performance metrics properly? Recommendation: Most processes already include guardrails so that large, sudden swings in performance metrics never result in massive budget increases. However, if yours does not, we recommend a maximum budget increase of 20% every 2 to 3 days. See the Incrementality Actionability article for more information.

Updated on: 28/07/2022