Meta Ads Structure for B2B SaaS – Acquisition Part

Recap from my presentation at Marketing Festival about my Meta ads structure. Today’s focus is on acquisition.

The Situation

You need to test different messages/propositions while targeting the same audience.

In other words, you need to regulate the volume on the ad creative level. And this is a very common case in B2B SaaS – you need to test various messages.

I will cover potential options for handling this situation, and at the end, I’ll present my approach.

#1 Having All Ads in One Ad Set

Meta doesn’t have an Ad rotation option like LinkedIn or Google Ads.

So when you run multiple ads, the Meta algorithm allocates most of the budget to the ad with higher engagement/soft conversions you are optimizing for.

And you simply won’t get enough data to evaluate which proposition works the best. Especially when you care about the actual product usage metrics, not just engagement metrics.

#2 Duplicated Ad Sets

Your budget is limited, and so the conversion volume is also limited.

Meta doesn’t have that relevant default targeting like LinkedIn, so you need to “narrow” the targeting through the conversion optimization goal.

Most projects simply can’t afford to have multiple ad sets. I usually have just one ad set for acquisition targeting and one ad set for testing.

Splitting the ads can ruin the optimization.

#3 Ad Set Scheduling

The same issue with low volume arises even when you schedule the ads to be running at different times. The only benefit is that you avoid auction overlap, but it is not the main issue here.

Even when using the A/B Testing feature, you need to have enough volume in each ad set for the conversion optimization. This feature ensures that the audiences will be evenly split and statistically comparable, but it does not share the data among the ad sets (source).

#4 Changing the Ads Each Week

Another solution could be using just one ad set. (Conversion data is not split among multiple ad sets.)

To ensure the ads get even volume, you could be running ads with proposition A each even week and then pausing it the odd week where ads for proposition B will be enabled. But pausing ads for 7 or more days might reset the learning since it can be considered a significant edit (source).

My Solution – Daily Custom Rules

Finally, what option works best for me is using Daily custom rules.

How it works and how to set it up

Define how much you want to allocate among the ads.

Let’s say you have a daily budget of $100 and two ads, and you want to allocate the budget evenly. For both ad A and ad B, it would be $50 ($100/2).

Once one of the ads reaches the limit, you need to pause it so the rest of the budget goes to the other ad.

There is a continuous custom rule that runs every hour or so. Define the conditions and action (in our case pausing the ad you selected and the today’s spend limit). Once the ad goes over the limit, it pauses the ad.

To make sure the ad will continue running the next day, just create another custom rule that would enable all ads every midnight.

You can have a different setup. In the slide below is a use case where you want to allocate the budget evenly on the product/message level, but you might have multiple ads for each of these messages. The custom rule can be unique for each ad.

Cons of this approach.

You probably have variants of ads for each proposition, but you don’t really care which of these ads run as long as it is delivering the same message. So even if you have some underperformer, it will get the same budget unless you change the rule.

My advice – Have a system/report to evaluate the ads on a regular basis. Go through the results and pause the ones that are less performing and reallocate the budget limit (condition for today’s spend).

Moreover, the best-performing ads (in the Meta view) are prioritized, so they can reach the limit already in the morning and so the results could be skewed if people’s behavior differs significantly based on the time of day.

Lastly, you need to remember to edit the custom rules when you want to pause any of the ads for good. Otherwise, the daily rule would enable it at midnight like the other ads.

The Results

The performance of the ads is stable. Daily pausing and re-enabling are not hurting the algorithm. The overall performance of the ad set can be worse since you might have less appealing ads with a poor CTR.

Surprisingly, Meta’s favorite ad can change over time as other ads (previously drowning kids) gather more data.

In the presentation that was part of the mini PPC camp program of Marketing Festival, I delved deeper into a different solution that I use for retargeting or smaller audiences. So stay tuned for the next article.

Bonus:

After the presentation, I was asked about the minimum conversion volume.

It depends on what your conversion setup is. What works best for me so far is to use conversions primarily to help the algorithm target the people in my category.

I try to combine several soft conversions to get as much data to the algorithm as possible so it knows it brought the right person. So the most frequent conversion in my account is when a visitor meets the demographics criteria (using IP address tools like Demandbase or Clearbit).

The official minimum recommended volume for conversion optimization is 50 per week, but from my experience, I try to get at least 130 conversions per week. Probably partly because of the conversion setup – some users could perform multiple conversions from a single ad click.

Practically, 130 weekly soft conversions translate to 600 monthly. To achieve this volume, I usually require 2000 visits (with an average CPC of $2.5), resulting in a monthly budget of $5000.

If anything here sparked a question, just ask me on LinkedIn 👇