Over the past 8 months, the team at Exmox has been iterating on hundreds of CPE campaigns and flows across all genres. A common thread amongst many of these campaigns is a fundamental misunderstanding of cost per engagement as a concept.
One of the first things you learn as a marketer, regardless of industry, is that there is no such thing as a free lunch. Behind those 600% ROAS months and viral game launches is a trail of unsustainable eCPIs, failed creatives, and games that were killed inside of their first few weeks. This is because experimentation and empirical data are the bedrock of effective campaigns.
Is CPE a fit?
When considering that CPE is designed to generate ongoing engagement in order to maximize returns, the need for experimentation could not be more evident. The core questions that need to be asked before pursuing CPE are:
Game Mechanics, is the game effectively monetizing and habituating users?
Have you structured a funnel by tracking multiple goals in your MMP?
Have points/levels/achievements in-game that indicate a high-quality user been identified?
The challenge with CPE comes down to identifying what flow results in the highest ROAS and best retention without burning through your budget. The process Exmox implements involve A/B testing a 7-day and 14-day goal, the goals should be in-game events, for example, reach level 20. Here is where patience comes into the mix if goals that take up to 7 and 14 days to reach respectively are being tested a wait time of minimum 2 weeks is needed to have an understanding of which approach is going to be scalable.
Many companies benchmark campaign viability off of day 7 ROAS causing the above approach to come under heavy scrutiny. Failure is a foregone conclusion if one is relying on a single metric to measure the success of multiple channels. Each channel needs to be taken case by case as the levers that impact performance vary. With CPE this could not be more true, in order to effectively gauge a CPE channels success consider the following:
Many CPE channels have a reward period this is the time in which users are being driven deeper in-game with micro rewards and other methods. Long term performance of a campaign can only be accurately measured after this period.
30-day performance, sometimes there is a need to adopt a “stop-start” approach with CPE in order to let the traffic and data mature. If a goal takes 14 days, run the campaign for a month, pause and evaluate the performance during the following month, this will provide a full picture without requiring a significant cash investment. Set attribution windows accordingly.
ROAS, keep in mind payments can occur before the goal is reached and, make sure ROAS is being calibrated to take this into account.
Monetization, if there is a hybrid model (in-app and ads) strong returns through ad events may be occurring, this is magnified by the low eCPIs typical of CPE.
The more goals tested the better, even more so if complementary goals to evaluate retention/engagement are set up alongside the paid goal.
Evaluating CPE requires a different mix of metrics. By taking into account the channel, goal targetted, genre, and in-game mechanics engagement campaigns become more predictable. The idea isn’t to abandon traditional metrics such as Day 7 ROAS, but rather adjust the mix of metrics used and adopt a channel-specific approach.