Creative Testing on a Budget. 


Video creative is one of the most important controls we have to attain profitable growth on Facebook. But finding a highly engaging ad that converts well requires resources and investment that not everyone has unlimited access to. 


Miri Growth is a performance marketing agency specialising in paid social UA for mobile apps and games. Based on our learnings from working with publishers small and large, we’re here to outline our tried-and-tested approach to creative testing on a budget.


Quality or quantity?


Our rule of thumb is to aim for high quality ideas and high quantity executions. One thing we often hear from smaller publishers is that they have an abundance of ideas, but lack the design resources to execute them all for UA purposes. Our solution is simple: spend time refining ideas to ensure they’re easy to digest from a user point of view, and source as many examples and assets as possible for the designer in order to enable easier execution. And a tip for any designers out there: we often find that our best performing ads aren’t necessarily the most polished or best-looking in terms of design – if in doubt, keep it quick and simple.


Iterate or explore new concepts?


Knowing when to iterate or drop a concept, or focus on only new creatives, can be challenging. These are the questions we ask ourselves when deciding what to brief each week:



If your KPIs for a new concept are miles away from your control ad, it’s probably worth dropping it. However, if it’s within 20%, there could be reason to iterate further.


If it’s been a while since you’ve found a new winning concept, it may be worthwhile thinking about what works and how you can make incremental improvements. On the other hand, if ad fatigue is kicking in rapidly and performance is poor, you may want to consider focusing efforts on finding a new well-performing concept to drastically improve results.


To A/B Test, or not to A/B Test?


A/B testing creative is core to our methodology at Miri, but that’s not to say it’s suitable for everyone. We value it so much because it allows us to guarantee delivery to all creatives we want to test and helps to reach conclusions with statistical significance. This is especially useful for comparing the click-through-rate (CTR), click-to-install rate (CTI), and hence the overall impression-to-install rate (ITI) against your control ad. However, depending on your individual KPIs, this can require significant spend on Facebook in order to obtain conclusive results. Additionally, some make the case that A/B testing is actually irrelevant, since Facebook works by picking a ‘winning’ ad and skewing most of the delivery to it, and guaranteeing the same delivery to all ads directly contradicts this.


But there is an alternative: target a country that’s both relevant and has the lowest CPMs, and group all new ads into one adset. In this case, we don’t include a control ad as Facebook is likely to favour it due to its historical performance. This can be a cost-effective method to achieve binary win/lose results for new creatives, but it won’t guarantee delivery to all ads. The ads that pick up the majority of spend and have a strong ITI can then be adopted into your main campaigns. 


Further reading


If you now feel confident producing and testing creatives on a tight budget, we recommend the following articles to take your UA creative one step further:


– We tested over 3,000 ads for 30 apps, here is what we learned (

– How to optimise UA creatives with iterations (


Zach is a Co-Founder at Miri Growth, a performance marketing agency specialising in paid social UA for mobile apps and games. They offer full campaign management with a strong focus on creative testing, using a team of in-house motion designers who work on 5 ads per app per week.