How can you exceed your KPI and still lose money, but then change your payment model and gain profit for both parties? We'll tell you about, using the example of our partnership with the game developer Goat Games.
Getting started
Goat Games (Guangzhou, China) came to Adsbalance to scale their audience for the Dragon Storm Fantasy mobile game and attract high-quality users.
Dragon Storm Fantasy is a mobile game with 3D graphics released by the developer of War and Magic strategy. The story is about a chosen hero who has to save the lands of Arcus. On the way, he engages in battles and destroys the main villain, including the support of assistants that contribute to a faster game experience.
The game was released in 2019 and was originally only available on iOS. As it developed, it became available on Android as well. It has an age rating of 17+.
Our buying team has been set targets with the following metrics:
KPI - Payback on day 7 - 38%
Geo - Europe, USA
There were no limits to the budget of the campaign, the main thing is efficiency and results. To begin with, we agreed to work according to the CPI model - payment for installing the application.
How the CPI model worked with Dragon Storm Fantasy
This model is quite popular for working with traffic among large partners. The customer only pays for the app installation, which looks the most optimal for a mobile game.
What results did we get?
Throughout the whole promotion period, we were working steadily in the plus for the client.
ROAS index was about + 30-70% for the volume of $50.000 per month. The best creatives showed ROAS of more than +100%.
The best audience came from these creatives:
We received a commission as a fixed amount per installation and tried to keep a balance between CPI and KPI on payback, for which we worked on creating the most effective and converting creatives and optimising the purchase. This allowed us to work in the black when we hit the set KPIs. But with this approach, we couldn't scale quickly.
What was the problem? As we scaled up, it became clear that it was getting more and more complicated to work with the CPI model. The objective of buying is not only to meet the KPI but also, if possible, to exceed it and optimise the client's costs. This has worked out very well this time.
However, with the current arrangements, it has been hard to work on this form of cooperation. If the buying exceeds the KPI, it gets paid the same amount as it is set to do. This means that the performers lose money. And the team's goal is to provide the game with quality traffic, make the client happy, and make money in the process.
Why was this the case? There is a difference between what the installer bought for and what they got in the end. Because the installation cost was fixed, the KPI did not show any real progress. It was not obvious what quality of lead was brought to the app.
In such a situation, the agency always has a choice. Either exactly meet one-to-one KPI, which in this case would mean having to pour in cheaper instals, so the figures would remain at the given level, but the quality of the traffic attracted would be lower. Or to work for over-performance, reducing your profits.
But we went the other way: we offered our client a new payment model - RevShare, thus getting the opportunity to earn more and giving the client the best quality traffic with exceeding their KPIs. We agreed to make a floating installation fee, which depends on our ROAS. Thus, it was possible to buy more expensive instals without any worries.
RevShare
RevShare is a percentage of the advertiser's profit that they get as a result of payments from app users attracted through an advertising campaign.
RevShare has pros and cons.
Advantages:
How the RevShare model worked with Dragon Storm Fantasy
RevShare is a truly mutually beneficial model. It's easier to run, scale and analyse results. This is achieved because when working with this scheme we can run multiple campaigns at once, for each of which we can view the results separately and, if necessary, disable it without affecting the other campaigns.
The latter is very important because several people are involved in a project at once. And if one employee can confuse the entire system when working with CPI, here the campaigns are autonomous and work more clearly. And therefore more effective.
This is how a real win-win occurs: a clear, transparent, and easy to manage the system for working with traffic, and an easy evaluate results. After we switched to RevShare, we tied the metric to ROI, which allowed us to work to our full potential, without limiting us in scaling.
Since April 2021, when we started working with this model, the following creatives have shown the best results:
In the end, by changing the model from CPI to RevShare, we shifted our focus from the number of instals to attracting a quality paying audience. This eventually gave a direct profit to the client, from which we earned a percentage.
What to Pay Attention to When Working with the RevShare Model
It is worth noting that the profit figure can only be predicted approximately. In contrast to CPI, where you get a fixed percentage of installations.
Remember that it is always possible to start a dialogue with the client, even after the agreements have been made. Especially if it's in the interest of both sides. The main thing is to show the right result of the advertising campaigns, as we did. As a result, we were able to increase our profits and continue to work on a mutually beneficial scheme.
Getting started
Goat Games (Guangzhou, China) came to Adsbalance to scale their audience for the Dragon Storm Fantasy mobile game and attract high-quality users.
Dragon Storm Fantasy is a mobile game with 3D graphics released by the developer of War and Magic strategy. The story is about a chosen hero who has to save the lands of Arcus. On the way, he engages in battles and destroys the main villain, including the support of assistants that contribute to a faster game experience.
The game was released in 2019 and was originally only available on iOS. As it developed, it became available on Android as well. It has an age rating of 17+.
Our buying team has been set targets with the following metrics:
KPI - Payback on day 7 - 38%
Geo - Europe, USA
There were no limits to the budget of the campaign, the main thing is efficiency and results. To begin with, we agreed to work according to the CPI model - payment for installing the application.
How the CPI model worked with Dragon Storm Fantasy
This model is quite popular for working with traffic among large partners. The customer only pays for the app installation, which looks the most optimal for a mobile game.
What results did we get?
Throughout the whole promotion period, we were working steadily in the plus for the client.
ROAS index was about + 30-70% for the volume of $50.000 per month. The best creatives showed ROAS of more than +100%.
The best audience came from these creatives:
We received a commission as a fixed amount per installation and tried to keep a balance between CPI and KPI on payback, for which we worked on creating the most effective and converting creatives and optimising the purchase. This allowed us to work in the black when we hit the set KPIs. But with this approach, we couldn't scale quickly.
What was the problem? As we scaled up, it became clear that it was getting more and more complicated to work with the CPI model. The objective of buying is not only to meet the KPI but also, if possible, to exceed it and optimise the client's costs. This has worked out very well this time.
However, with the current arrangements, it has been hard to work on this form of cooperation. If the buying exceeds the KPI, it gets paid the same amount as it is set to do. This means that the performers lose money. And the team's goal is to provide the game with quality traffic, make the client happy, and make money in the process.
Why was this the case? There is a difference between what the installer bought for and what they got in the end. Because the installation cost was fixed, the KPI did not show any real progress. It was not obvious what quality of lead was brought to the app.
In such a situation, the agency always has a choice. Either exactly meet one-to-one KPI, which in this case would mean having to pour in cheaper instals, so the figures would remain at the given level, but the quality of the traffic attracted would be lower. Or to work for over-performance, reducing your profits.
But we went the other way: we offered our client a new payment model - RevShare, thus getting the opportunity to earn more and giving the client the best quality traffic with exceeding their KPIs. We agreed to make a floating installation fee, which depends on our ROAS. Thus, it was possible to buy more expensive instals without any worries.
RevShare
RevShare is a percentage of the advertiser's profit that they get as a result of payments from app users attracted through an advertising campaign.
RevShare has pros and cons.
Advantages:
- stable income - you'll receive royalties for as long as the user pays within the app;
- all the focus is on results, not on calculating a model that will allow everyone to come out in the black;
- higher ROI, which has no limits;
- direct correlation between ROI and quality of traffic;
- no link to price;
- possibility to buy quality traffic.
- tests are more expensive compared to the CPI model;
- it takes time to get the result of a campaign;
- no exact test result is available right away, only indirect leading indicators such as CPI or CTR are available;
- users may buy something in the game only after a certain period - it takes longer to get the percentage;
- it is quite difficult to analyse users;
- it requires a bigger budget for the test;
- requires a deep dive into analytics.
How the RevShare model worked with Dragon Storm Fantasy
RevShare is a truly mutually beneficial model. It's easier to run, scale and analyse results. This is achieved because when working with this scheme we can run multiple campaigns at once, for each of which we can view the results separately and, if necessary, disable it without affecting the other campaigns.
The latter is very important because several people are involved in a project at once. And if one employee can confuse the entire system when working with CPI, here the campaigns are autonomous and work more clearly. And therefore more effective.
This is how a real win-win occurs: a clear, transparent, and easy to manage the system for working with traffic, and an easy evaluate results. After we switched to RevShare, we tied the metric to ROI, which allowed us to work to our full potential, without limiting us in scaling.
Since April 2021, when we started working with this model, the following creatives have shown the best results:
In the end, by changing the model from CPI to RevShare, we shifted our focus from the number of instals to attracting a quality paying audience. This eventually gave a direct profit to the client, from which we earned a percentage.
What to Pay Attention to When Working with the RevShare Model
It is worth noting that the profit figure can only be predicted approximately. In contrast to CPI, where you get a fixed percentage of installations.
Remember that it is always possible to start a dialogue with the client, even after the agreements have been made. Especially if it's in the interest of both sides. The main thing is to show the right result of the advertising campaigns, as we did. As a result, we were able to increase our profits and continue to work on a mutually beneficial scheme.
Last edited by a moderator:
