Use Power Editor.
Its ease of use will can allow you to build and set multiple Facebook ads campaigns live all at once. It also simplifies duplications to take the headache of repetitive tasks out of the way.

Consider the size of your audiences.
Your ability to stay relevant (to an extent) is extremely important when looking at Facebook costs. Use the tips below to build the best audiences:
Start with re-marketing
Build lookalikes off of re-marketing
Combine these two or use interests, behaviors, and demographics to construct a relevant audience toward your previously defined objective

Focus on the value proposition of your promotion when building Facebook audiences manually (using interest/ behaviors). You can then work backward to target the people that it would benefit the most highly. Try to hit a sweet spot with the audience—reach and costs will be directly correlated to audience size and, ultimately, the relevance and quality of your ad.

A/B test ad sets and ads using multiple targeting combinations and ad copy – optional use of multiple images and messaging. After they run for a period of time you can go into each and optimize by allocating budget to higher performing ads and cutting out the lower performers.

Identify the placements in each ad set that are driving the highest, best quality results. You can break the ad set up further (using power editor) and have individual budgets for specific placements. For example, if the desktop newsfeed is driving more leads than the app network and mobile newsfeed. You can break out that placement and devote all of your budget to the desktop newsfeed. Facebook will automatically optimize and pay more attention to higher performing placements. This can be deceiving; for example, in a web clicks campaign you could drive high traffic from the app network, but in reality these clicks are accidental. Facebook will think that particular placement is the highest performer and shift to spend more money there.

No Comments

Be the first to start a conversation

Leave a Reply

  • (will not be published)