I have created and tested over 400 video ads for 33 different games and prototypes and I wrote over 70 marketing data analyses. I’ve been part of a team that worked hard to create epic experiments to improve metrics. And on top of that, I’ve really taken the time to dive deep into the skills I want to master. And in doing so, I’ve collected a bunch of learnings that I’m now ready to share.

How to design a winning creative

We all know that most of our creatives fail. Out of all the creatives I’ve tested this year, only 1-3% fit into the winning category. And I’m not the first one to figure this out. Big publishers and marketing agencies have shared findings pretty similar to mine. This means that, to find that special unicorn ad, you have to test a lot of ideas. Sadly, Creatives can quickly lose their effectiveness due to short attention spans, quick changes in the market and ad-fatigue. This means that if viewers see the same or similar ads multiple times, they won’t engage with it anymore. That’s why you need to keep moving, even if you have been so lucky to find your unicorn ad. You can do this by continuing to test completely new ideas and by iterating on the best performing ads you already have.

Types of creative ideas

I personally divide new creative concepts into 3 categories: Experiments, research based ideas and iterations. I’ll start with what they should all have in common:

  • All ideas should have a theory of why they could work. And you should have reasonable confidence of the theory and excitement about the idea.
  • All ideas should be able to be finished within a reasonable timeframe. In our marketing team we don’t really work with tight deadlines, but I personally like to only work on ideas I can have done within the next release date. I’ve found that if I commit to a bigger project it always has to be lower priority. And lets be honest, no one has time for low priority tasks.
Now for the differences:
  • Experimental ideas follow the least amount of rules. They can come from anywhere and haven’t really been proven to work yet. They are mostly inspired and accepted within a team because people get excited thinking and talking about it. Experimental ideas might be the highest risk, but they’ve also got the opportunity to get ahead of a trend or be the first in a field.
  • Research based ideas come with the confidence that an idea has either worked in the past, or is currently working for one of your other games or for competitors. They can also come from current trends and events or viral happenings on the internet.
  • Iterations are ideas to improve on, or test different variants of, a creative that has already been tested. It’s important to only change one thing per ad so it’s easier to draw conclusions from a test. If you change too many things at the same time it will be hard to determine which change had the most impact on your data.

Ideas that have proven to work well in 2021

So what are the kind of ideas that have worked well for mobile game ads in 2021 and will most likely continue to work well for 2022? I’ve compiled a list:

Clear gameplay footage

You only have 15-30 seconds to convince the viewer to play your game, use this time wisely. Complex gameplay might look really cool to you because you’ve been playing this game for hours, but that doesn’t mean a fresh audience who has never seen your game before is going to understand what is going on. Make sure to remove anything that is not helping to explain the game. And add anything that could help with this.

Gameplay that includes choices

Games that have simple A or B choices have been skyrocketing in the App Stores this year. It wouldn’t hurt to use this same principle in your ads too. To do this successful you have to give the viewer a meaningful choice where the outcomes of either choice has a big impact.

Fail gameplay footage

Have you ever seen anyone play a game for the first time and totally failing at it? Don’t you want to take over and show them how its done? This is the type of creative that keeps performing very well and understandably so. The idea works best if the gameplay is easy to understand, and the player in the ad makes obvious wrong choices while the right choice is right there on the screen.

Gameplay that evokes emotions

Humans are feeling creatures and this means that we’re more likely to respond to something if it emotionally affects us. This could be thinking that something is funny, feeling accomplished and powerful, but frustration, anger and fear are also really strong motivators to engage with content. There’s a reason horror games are still popular. If you know how to use emotions in your ad well, you might have a winner on your hands!

Gameplay with big contrasts

You can take this advice in the biggest sense of the word, but examples of big contrasts that have worked well this year are: big vs small, bright vs dark, slow vs fast, chaos vs calm, noob vs pro.

How to learn from testing so many creatives

It can be pretty overwhelming to test so many different ideas. That’s why it’s important to keep things organised. This can mean different things to different people. The marketing managers before me loved spreadsheets, but I found them difficult to read and, without any context, hard to learn from. I decided I needed something more visual. I started documenting data outside of spreadsheets, and included color coding for different benchmarks, and screenshots to visualise what I was talking about. As I was working on this structure I often asked and received feedback from different people in different positions within the company. They shared with me things they found most useful. I suggest you find out what your team needs in order to get the most out of the data you’re collecting.

Don’t forget to check your own results with what your competitors are doing! Most ideas aren’t new, and changes are that people have already spent money testing the exact same thing. Doing this research can safe you a lot of time and money. Speaking of doing your research, did you recognize all mobile games in the header?

Happy testing!