I was recently reviewing a few of my client’s campaigns when I came across some common mistakes present across campaigns. Thus I thought that I should share them here to hear alternate suggestions and potentially get a discussion going!

Common mistake #1: Placing multiple audiences together in one ad set 

In this case, my client was placing multiple lookalike audiences in one single ad set as shown below.

Why it matters

Placing multiple audiences (lookalike or otherwise) in one ad set prevents you from identifying which specific group of audience drives the best results. And why should you aim to narrow down your audience? Doing so ensures that you won’t be spending unnecessarily on audiences that aren’t driving results.

How to solve it

Shortlist several audiences and place them in separate ad sets under a single campaign (as shown below). This way, when you’re viewing performance at the ad-set level, you’ll able to identify the better performing audiences as they aren’t all placed together in a single ad set as before. If desired, you could run a split test to be sure which specific audience drives the best results. 

As a rule of thumb for shortlisting lookalike audiences, the source audience which the lookalike is created should correspond to the objective that you’re optimising for. For example if you’re optimising for purchases, my hypothesis would be that the lookalike based on a source of your top spenders would perform best. However, a source audience of people who have added items to cart is well worth a shot too, especially if the former audience pool is too small. 

Common mistake #2: Repeating your audiences across campaigns or ad sets

Not only was the client placing multiple lookalikes in one ad set, she was doing so with these same lookalikes (e.g. Lookalike #1, #2, #3) across multiple campaigns*, as shown in a simplified illustration below:

Why it matters

Repeating your audiences across campaigns puts your ads up for competition against one another during an ad auction, inadvertently driving costs up, and thereby reducing the returns on your ads.

While the client didn’t make the mistake of repeating audiences across the ad sets, doing so may lead to poor delivery of ads. The same applies to when there are overlapping audiences across ad sets under a campaign, as Facebook will ‘enter the one with the best performance history and prevent the others from competing to get shown.’

How to solve it

Refrain from using the same audiences across campaigns (as shown above) or ad sets (shown below). 

In a nutshell, all ad sets (regardless of whether they are under the same campaign or in other campaigns) should have unique, non overlapping audiences. If you have several ads that you would like to show to a single audience, create multiple ads under one ad set as shown below. 

Common mistake #3: Not having exclusive campaign(s) for retargeting

In this case, my client placed the retargeted audience along with potentially new prospects (the lookalike audiences) in a single ad set - which meant serving the same ads to these two groups of audiences.

Why it matters

Retargeted prospects and new prospects differ in terms of the level of understanding, interest and intent in the brand and/or product. Placing them in a single ad set and thus serving them the same ads risks having the ads being irrelevant, and thus ineffective. 

How to solve it

Create separate campaign(s) for your retargeting efforts. Even within a retargeted audience, the differences are still significant in terms of their interest (e.g. browsing different pages) and intent (e.g. viewing the home page vs dropping off during checkout process). Further slicing your retargeted audience and customizing your ads accordingly allows for your ads to be more relevant and effective.

In terms of having different interests where users browse different pages of your site, the ad would work better if it retargeted specific (key) pages of the site and tailored its messaging and creative accordingly. For e-commerce businesses, this means utilising Facebook’s dynamic ads to show people the specific products they have viewed / added to cart but not purchased.

In terms of difference in level of intent, you might want to consider retargeting those of a higher intent (dropping off during checkout process as opposed to merely viewing a page) with higher bids and exclusive offers as opposed to doing the same across the board.

Summary

In summary, here are the 3 common mistakes and what you should do instead:

Mistake #1: Placing multiple audiences together in one ad set 

What you should do: Place only a single audience in one ad set as it allows for identification of the better-performing audience  

Mistake #2: Repeating audiences across campaigns or ad sets

What you should do: Avoid using the same audience across campaigns and ad sets as doing so will result in higher costs and poor delivery respectively

Mistake #3: Not having exclusive campaign(s) for retargeting

What you should do: Create exclusive campaigns for retargeting and further segment your retargeted audience

Conclusion

This is my take on the issues I’ve picked up - what’s yours? There is no ‘one ultimate way’ to iterate an ad, which is what makes the iteration process exciting for me. 

Also I feel like I’ve barely scratched the surface with this article, especially with the topic of best practices for source and lookalike audiences. What would you want to see me discuss next? Leave your comments down below, I’d love to hear them! (: