Finding Creative Signals for Success
Why you can never know why a creative worked
The reality is that nobody knows. Just look at your campaigns and look at the creatives that worked - can you try and articulate why these creatives worked?
Nope.
You might be able to articulate some reasoning - but there is no way to know for sure.
You can go through a few tests to see this in the real world:
- Does the creative you think will perform before uploading it are the actual ones performing? Can you predict it consistently?
- Were you able to replicate success over time, consistently?
It's not impossible to know because your expertise and experience are not valuable, you are not even able to know for sure who saw that creative as the audience data itself is unknown - you just see its performance.
In this case, deriving expertise-based insights aren't resulting in massive wins.
To say that a certain creative performed because there was a blue hat, or this character, or that mechanic is simply a guess.
But a guess reality shows us is not yielding consistent performance.
The field of creative tagging for various assets has existed for a while without mass success.
Don't try to fight it
The most successful UA and creative teams right now - that have chart-topping games that make a good chunk of the entire mobile game market globally, have something in common.
Massive exploration of creatives - sometimes uploading thousands and thousands of creatives each month for their games. This allows them to provide the ad networks with exactly what they need to find success - massive experimentation and exploration.
They still don't try to chase a process that relies on some genius moment of ideation. They simply allow for a process that systematically has a better chance than anything else to find their game valuable audiences.
We view it as a search problem - not an ideation game
We view this here at Sett, as a search problem - a problem of searching for these winners in an almost infinite space of possibilities with an intelligent system that can:
- Analyze massive amounts of data and find patterns that we humans simply can't
- Form hypotheses, generate new video and playable UA creatives with significantly different concepts
- Deploy them and monitor performance simultaneously of massive amounts of creatives
There is no secret sauce.
Heck, even J.K. Rowling couldn't find success after Harry Potter - how are we expected to find the next winning creative after one success we can't even explain? There's a better way - stay tuned for more.
