This is a summary of the most important best practices that any advertisers should test and which we recommend following. Most of them are fairly easy to set up and implementing them should have a positive impact on your performance. They have also been seen working repeatedly in practice with customers that we have worked with in different industries. It is naturally suggested that each advertiser tests these by themselves, especially if following them is a drastic change compared to your current way of running Facebook advertising. After testing these guidelines in practice, you can start testing new ideas to determine what improves performance in your context.
1. Follow a full funnel approach
This means that you have the campaign objective, audience, and creatives tailored according to each funnel step.
Full funnel in Direct Response advertising
Why should I care? The answer to this depends on the advertiser, particularly the size of their budget and their target audiences' awareness of their brand. For example, established and well-known brands might benefit less from brand awareness or traffic campaigns. In the cases of less-known brands, or brands that have a limited amount of organic traffic, we have seen success by combining top of the funnel campaigns with prospecting and retargeting campaigns.
How can I test or apply this? With a smaller budget, you might not be able to build all the funnel steps before you have reached a certain scale. The typical starting point is simply having a prospecting campaign and targeting towards your final goal, such as purchases, with a creative focused on the product itself.
Here are a few examples of tests that you can run to determine which funnel steps work for you with lift studies:
- Can brand awareness campaigns bring an incremental lift to my Facebook advertising?
- Can adding a traffic campaign help in increasing the scale of my Facebook advertising?
- Does retargeting bring an incremental lift to my Facebook advertising?
2. Create an appropriate audience structure
The audience structure refers to how audiences are split into ad sets under each campaign. The structure is typically maintained for a longer time.
Why should I care? Often, advertisers split or duplicate audiences unnecessarily, resulting in worse performance. More specifically, this can be due to the following:
- Splitting audiences: the audiences may become too small, possibly resulting in ad sets struggling with collecting enough conversions on a weekly level for Facebook to know whom to serve the ads to.
- Duplicating audiences: “Having overlapping audiences is not necessarily a bad thing, but it can lead to poor delivery of your ad sets. This is because when ad sets from the same advertiser end up in the same auction (meaning they're targeting similar audiences), we enter the one with the best performance history and prevent the others from competing to get shown.” Read more about overlapping audiences
Structuring audiences according to the funnel steps and maintaining them over a longer period of time includes multiple benefits:
- Campaigns are easier to manage and analyze;
- Audience performance history is collected to one ad set, making optimization and analysis more reliable; and
- Optimization triggers work better. Improving results requires iterative changes over 5+ days. A fixed structure works better.
How can I test or apply this? Here are also typical tests to determine which audiences you should use and how to structure them:
- Broad audiences vs. interest-based audiences
- Different seed audiences for lookalike audiences
3. Select the right bidding type and goal
Why should I care? Selecting the wrong bidding type can hinder scaling your Facebook advertising, if it becomes a limiting factor i.e. it prevents you from gaining more conversions.
How can I test or apply this? When it comes to the bidding goal, bidding towards the event lowest in the funnel (e.g. a purchase or subscription to an app) gives typically the best results.
Changing bid goal resets learnings so we don't recommend first bidding higher in the funnel and changing to a lower event when there are more conversions. If there are no enough events, you go for an upper funnel event. Changing the bid goal once does not trash the performance, but a strategy where that is done systematically for all the campaigns is not recommended. If unclear – test with an ad study.
4. Run enough ads per ad set
Why should I care? Running only a single ad in your ad set means that you have only a single version of a creative for trying to convince a user to take the action you want. Put another way, you are placing your bet on a single ad. However, every person in the audience is different and different things works for different people. If you have more ads targeting the same audience, you increase your chances of converting people with content that is more relevant for them.
How can I test or apply this? We recommend you to generally have multiple ads per ad set for Facebook to optimize which ad works best for each individual. However, having too many creatives is wasteful, since the Facebook delivery algorithm decides which ads to show relatively fast.
Facebook recommends having a maximum of 6 ads per ad set. We recommend an absolute maximum of 10. In special cases, where you have tested that a creative performs well and there is no danger of ad fatigue, you may go with just one ad.
5. Run ads with enough of variety
Why should I care? To reiterate the previous point, different things work for different people. Facebook advertising is also becoming more competitive as advertisers put more focus on the creative. Hence, it makes sense to run different kinds of creative types.
How can I test or apply this? Here are example tests on how you can see, which creative type, or which mix of creatives, works best for your target audience:
- Creative split test between multiple creative types
- Split test with a single creative type vs. a mix of creative types
6. Selecting the right audiences (right target audience, right audience size)
Why should I care?
- Right target audience: Facebook can usually find the right people within an audience for you to advertise to, so using overly granular audiences is unnecessary. However, if your product is more niche, you might be wasting money on people that are not relevant to your business.
- Right audience size: Too small audiences can become saturated quickly. On the other hand, targeting too large audiences may be a waste of money if your product is relevant only for a part of the audience.
How can I test or apply this?
- Right target audience: If your product is applicable for the majority of the population (e.g. T-shirts), broader audiences should work for you. With more niche products (e.g. high-end luxury watches for women), you need to be more specific.
- Right audience size: First think of, who you want to reach, as highlighted above, and what do you want to achieve with your advertising. if you want to reach everyone in the population, your audience size will be large (eg. all Facebook users in the US). if you need to be more specific, your audience will be narrower. Optimal prospecting audience size is mostly in relation to advertisers budget (try to reach some 10% of the audience to give Facebook room to optimize who to target).