When advertisers create their campaigns, they tend to fall for the idea that more campaigns equal better performance. Overcomplicated campaigns that are impossible to follow and new campaigns that are launched daily can harm your performance more than it can help. We all want the secret sauce and the professional strategies that the big advertisers use. And it’s not just business owners who run their campaigns – ad agencies and the entire industry should be held responsible. If you are paying someone to run your ads, understand that oftentimes, making overcomplicated campaigns is part of their business. They want what they’re doing to look extremely advanced. You wouldn’t settle if you just saw that they created a single campaign, even if it’s based on the best practices. You would want to see many new ideas being tested and campaigns created. The idea of creating tons of different ads to boost performance is fundamentally wrong and works against the best practices of advertising on Facebook. Fewer ads will lead to better performance. Let me explain why.
Without going into all the details of how the Facebook optimization algorithms work, let me break it down. Facebook ads are designed to work based on pixel data and rely heavily on past data. This means that the targeting, delivery, and performance optimization will be dictated by the data from today, yesterday, and the past 180 days. Learning is stored in two main areas. The first is on the ad set level, which is the sublevel of the campaign. When your ad set starts getting conversions, conversions are stored in that ad set history and that data is used for learning and reference to find similar people who might be most relevant (lookalikes).
The second area of learning is your Facebook pixel that’s installed on your website and tracks every action. If you have the pixel installed on your site currently, Facebook gets access to every piece of data on your website even if you don’t presently run ads. They use that information and match it with users on their platform to understand and classify users based on the actions they take and the interests they might have. When you run ads, Facebook combines the ad set learning and the pixel data for the best performance. Up until this day, cross-campaign learning doesn’t exist or at least doesn’t create a big impact. If you have hundreds of different ad sets running, they each have their learning system and they share very little data between each other. They will get information from the pixel, but your ad sets probably won’t have a lot of data to work with.
Some Facebook marketing experts use the term “winners” for ad sets with more than 50 conversions per week and use the term “beginners” for ad sets with few conversions. If you have an ad set that gets 50 conversions per week, imagine how much data it already has, especially if it has been running for a few weeks or months. Top that off with data sharing from your website pixel, and you have a monster, where delivery is on point for almost every dollar spent and performance is strong and steady. Then, take an ad set with a few conversions per week, and you will find an ad that delivers unsteady performance because there is very little data for the algorithm to work with.
Then why doesn’t Facebook simply use cross-campaign learning to let you leverage all the data from your ad account? That’s a good question and there is no official answer, but I have a pretty good idea why. Campaigns are meant to be the shell for different categories, topics, or creative ideas, i.e. one campaign for a type of product, one campaign for this creative concept, or this goal, etc. If my theory is correct, then cross-campaign learning might be a little less relevant because one campaign might target a specific audience, while another might target a different set of products and so on.
What’s my point here? If cross-campaign learning doesn’t exist and the amount of data per ad set matters, you need to create fewer campaigns and focus on growing your existing ones. When I talk to other experienced marketers and discuss my strategy methods, they are always shocked by the number of campaigns I am running on large ad accounts and how simple I keep my ad structure. Even with ad accounts that spend millions of dollars per month, I keep the number of running ads as low as possible to ensure that I leverage the targeting optimization in the most effective way possible instead of having it work against my goals. I never create duplicate campaigns that target the same people. The only time I create more campaigns for the same audience is only to showcase new products/offers or fresh creatives.
Creating the “Snowball Effect.”
Like a snowball that rolls down the snowy mountain and becomes larger as it rolls, the same thing happens with your ads. The more data your ads have, the better they will perform. Your goal is to get the ad set from zero conversions to 50 conversions per week as fast as possible. When you have an ad set that has 50 conversions per week, performance will be steady and you will be able to scale your ads to big numbers. The goal here is instead of growing the number of active campaigns we have, we should focus on growing the existing ones.
Like any campaign, opportunities are not infinite and at some point, performance will start decreasing and you will need to work on a new campaign with fresh creatives/offers.
When Should you Create New Campaigns?
I like to think of the word campaigns as a common concept. What do all the ads in the campaign have in common? They all share a specific concept. Either they all have the same format (video, image, carousel ad, etc.), or they all target the same audience or a specific promotion. Within your new campaign, you would create different variations of that creativity to a/b test the image and copy, but all creatives would follow the same concept. Refrain from duplicating and creating the same campaigns if nothing new is being introduced.
To understand the best ad structure, it’s important to understand the logic behind the optimization algorithm. Facebook ads work based on previous data to optimize your campaign’s performance. This means the more data, the better the results. Data is stored on an ad set level and cross-campaign learning doesn’t have an impact. Your goal as an advertiser is to have a small number of running ad sets with large budgets and many conversions to achieve the best performance. New campaigns should be grouped by similar concepts and the creation of new campaigns should be done when new creatives, products or offers are introduced.