How To Test Ads and Improve Your ROAS by 97%

Hey friend,

How was your weekend? I spent my weekend snowboarding… it’s a great way to not think about advertising, although the occasional ad idea does pop into my head on the chairlift! 🏂

Right - let’s dive in. One of the questions you all have been asking me is ‘how do you test ad creative?’

After working on over 100 ad accounts and spending literally millions every month, I refined an approach that works well without being complicated.

This week I am going to break down for you exactly how I do it. This is the system can literally turn an ad accounts around - such as the beauty brand we improved performance by 97%.

I think a lot of people over-complicate testing - it’s not easy to find winning ads, but your testing set-up can be simple.

Embrace the "No More Winners" Mindset

Ditch the idea of one single ad carrying all the weight. It's not just about finding "winners," but cultivating a diverse portfolio of ads that all contribute to the overall success of your digital advertising campaigns.

Think of it like an investment portfolio – diversifying minimises risk and increases the potential for success.

The goal is to have multiple assets performing well, each catering to different segments of your target market or stage of the customer journey.

You can achieve this through creative exploration; experimenting with ad formats, messaging, and visuals.

Define Your KPIs

Before you start or launch any test - you need to clearly define your control metrics (what are you actually measuring and how) based on your current goals and the seasonality of your business.

Also, understand what a good benchmark is. For instance, your target CPA during the holidays will be drastically different to what you see in January or February.

My Simple Structure For Testing Ads

The typical testing structure I use is to test ad concepts in their own Ad-Sets, testing them against winning concepts in a CBO (that stands for ‘Campaign Budget Optimisation’ and is a campaign where the budget is set at the campaign level, with Ad-Set budget decisions made by the algorithm).

So what this looks like is one main Ad Campaign, within that Campaign I have one ‘winners’ Ad-Set where we graduate all winning creative. This acts as the control.

We launch each test in a sole ad-set. Each Ad-Set represents a ‘test’ or hypothesis and usually is a ‘big swing’ with micro variations of ads tested inside the same Ad Set.

Then every time we want to launch a new ad test - this might be a particular marketing angle like ‘our diapers are eco friendly’ or it might be a new concept that communicates an angle we already know works eg ‘3 reasons why our diapers are eco friendly’ vs a winning version of that angle that has worked before.

Each Ad-Set has one single concept with some micro-variations that we want to test - we might do one different 3-second hooks or we might have different music or a shorter vs longer version. The angle and concept are still the same, but we want to fine-tune different versions to see what performs. The micro-variants sit together in one ad set. I launch each ad with a couple of headlines and primary text versions.

In a smaller ad account we might try to launch one new test every week - in a bigger account or with more budget we would launch more.

Once the test is launched we leave it up to Meta to find the winners - and will graduate an ad into the winning Ad-Set only when it has proven itself over a sustained period.

Real life example

Sustainable diaper brand

• Ad set test one: ‘3 reasons why’ - sustainable angle. 3X videos, 2X headlines 2X copy.

• Ad set two: ‘Agitate Painpoint solution’ - sustainable angle. 3X videos, 2X headlines 2X copy.

• Ad set three: ‘TikTok Comment Response Ad’ - cost saving angle. 3X videos, 2X headlines 2X copy.

• Ad set four: Past winning ads.

Budget: $200 per day set at campaign level

Targeting: Broad with age restriction, exclude 14 day buyers.

Don't Discard – Repurpose and Optimize

While we let Meta do the optimisation and allocate the budget to the best concepts, we can watch the data and decide what to do next.

If a certain angle is giving good results, we can test new creative ways to communicate that same angle with a new ad. Conversely, if a certain style of ad works well, we can try that style to communicate new marketing angles.

We can also use the data to identify ads that show promise, but may need some tweaks. An ad might have a great ‘hook rate’ but need some improvements after that. Equally another ad may be performing quite well, but not hooking people as well as other ads. In this case we’d try some new hooks on the same base unit.

Testing never truly stops. It's an evolving process of analysis, optimisation, and constantly seeking that sweet spot between engagement and conversions. Even "losing" campaigns can yield valuable data points that help you refine future creative.

While seeking groundbreaking wins is fantastic, don't underestimate the impact of consistent incremental gains. Are your new ads incrementally improving over past social campaigns?

Creative Testing During Low Performance

It can be scary when performance drops - but if your winning ads are no longer winning, it is time to double down on testing. Here are some tips for turning things around when they go safe.

• Instead of scaling back testing, actually do the opposite. This can sometimes help overcome issues with data loss, especially on a platform like TikTok Ads.

• Increase your testing budget to around 40-50%. It's a good idea to have separate campaigns for testing so that it doesn't negatively impact the performance of your main campaigns.

• Think of it as needing to gather more "data" during these periods to see what works. Embrace bold, out-of-the-box ideas to see what sticks.

• Remember, it's easy to fall into the trap of over-analyzing minor details when things are going well, but bigger swings are required during slumps.

Staying competitive within Meta in 2024 requires continuous learning and adaptation, but having a simple structure like this to launch new creative and harness the algorithm can allow you to focus on what is most important - coming up with new ad ideas.

Always keep the bigger picture in mind – building brand equity and delivering genuine value through content that resonates with and captivates your audience. So get out there and make it happen.

Your future unicorn ad creative is waiting to be discovered, but in the meantime, you can create some reliable workhorses to bring you solid returns.

Have a great week!

Jessie

Ps. If this was a lot to take in and you feel like you might want my help - book a call and let’s have a chat! I’d love to meet you!