What Sett is building → the winner-search loop
We're tackling one of the toughest challenges in mobile game growth: finding winner creatives and succeeding in UA.
The advertising environment for mobile games has changed dramatically over the past few years.
From the reliance on what some called the most comprehensive database on humans - meaning Facebook pre-iOS 14 and ATT, to a couple of years of shifting forces, to the rise of SDK networks - with Applovin leading the charge.
This increased the importance of creatives as a factor in succeeding in user acquisition for mobile games. If in the past creatives played a smaller part in the success of campaigns on the back of highly targeted campaigns based on user-level data (where the targeting was so good the creatives had less "weight" to carry), today they are pretty much the last frontier.
With everything else being handled by the black boxes of the SDK networks, such as Applovin - these machines basically evolved to extremely efficient systems that find audiences for creatives.
These machines find audiences for certain creatives, and evaluate whether that audience is responding to the ad and then in turn has a high enough percentage of high LTV users, either in the form of highly engaged users that drive IAA revenues, or paying users that drive IAP revenues.
How are things today?
According to what we see, roughly 2%-3% of creatives ever manage to get a significant share of spend, and hopefully become those rare winner creatives, or hero creatives as called by some.
So many games reach their prelaunch or even launch phase without being able to find a valuable audience to the game. A big part of that is that these games failed to find a formula to market their games.
Teams are struggling like crazy to find winning creatives. Unpacking it from an old school marketing perspective simply means they were not able to find the right way to market their game. They were unable to find a creative that users responded to and also led to them loving and having fun within the game.
Where's the problem?
One of the problems is that it appears utterly random to us humans.
Entertainment can't be predicted
We don't f*cking know. Nobody knows.
There are tens of billions of dollars a year being spent by entertainment companies every year on trying to land on a hit.
And yet, for every global hit, there are many, many hundreds of projects that failed.
It's not just true for games; it's true for any form of entertainment. Trying to predict what people would love is a fool's errand. It's a numbers game and everyone knows it.
The issue is that because the cost of feeling out an idea is so high, the entertainment industry, as well as the gaming industry in particular are trying their best in placing the right bets.
Focus groups that mean nothing, massive consumer intelligence departments, and more competitive intelligence insights than anyone even cares to read.
If it worked, we would already see a company with geniuses putting out hit after hit without stopping. But yet, we live in a world that for at least the past 28 years, there is only one GTA.
It's f*cking hard to solve.
In the realms of user acquisition teams and creative teams, that problem is no different. Relying on hunch, intuition, and genius game, UA, and creative folks, still results in just 2%-3% of creatives ever finding success.
The winner-search loop and how Sett views it
We call this entire process the winner-search-loop.
Today it looks something like this:
Creative and UA Research
You try to come up with the best few ideas you can in order to prioritize the lengthy and costly creative production process.
Past campaign data is analyzed by a human team that tries to mine for insights and any type of idea that'll make the next round of ideation better.
The team researches, reads through past data, looks at competitors and trends, finds references, and prepares briefs.
Creative Ideation
You try to come up with the best few ideas you can in order to prioritize the lengthy and costly creative production process. You research, read through past data, look at competitors and trends, find references, and prepare briefs.
Creative Production
You pass those briefs to artists and developers for playable ads who need to interpret your brief and hopefully create something that distills the essence of the idea. This process usually takes weeks and sometimes even more when it comes to complex formats such as playables.
Creative Deployment
These creatives get uploaded, those that get approved at least, to various ad networks and we're off to the races.
The ad networks run internally dozens or hundreds of tests trying to find pockets of valuable audiences for the creatives, most times fail and the creative is dead in the water.
The result is one out of 30-50 creatives that succeeds.
Sett Agentic AI Platform for UA Creative Performance: The system of winning
We believe that in order to win you need massive scale, one that we believe is achievable through our agentic AI system that tackles the entire winner search loop.
We're building a platform that is:
Generating diverse, high-quality UA videos and playables
Creative diversity is critical because winning creatives exist somewhere in a vast, nearly infinite space of possibilities.
We believe that the best winner-search process is one driven by massive testing of that space in scale, by generating different and varied concepts for your game representing very different approaches to how to market the game.
This part of our system unlocks scale by generating hundreds of new concepts to test in a way that would not be feasible to any human team.
Building a robust data infrastructure
The next phase involves collecting meaningful signals from deployed creatives and other sources to enable intelligent, data-driven decisions.
Instead of relying on human interpretation of limited data, enabling intelligent AI-driven systems and reinforcement learning to find patterns in performance data to lead the winner-search to zoom in on more promising areas of that infinite creative possibility space.
Automating the entire loop
The ultimate goal is to have the platform autonomously research data, generate ideas, create content, deploy it to campaigns, and learn from results - creating a continuous, intelligent search for winning creatives.
Wrap Up
In short, we're building a system that will allow any UA team to have the most robust and effective way of finding winning creatives and turn their games into hits.
We started with creative production and supplying you with the ammunition needed for success and we're quickly pursuing the other avenues to create the most powerful game success engine we can imagine.
Stay tuned, if you don't want to miss out - sign up for our waitlist!
