Skip to main content

Optimizing & Scaling Your Facebook Campaigns – Step by Step

A lot of people have been asking me, how the hell was I able to get to XXXX per day with TEESPRING in a manner of months; because when they tried and launched shirts they failed.

My trick is simple – once I find my unicorn I work real fast to scale that one unicorn as fast as I can. All while, my team is continously launching new shirts every single day. We have this saying “No idea is a bad idea” so we try every single stupid t-shirt idea we can think of, and many times they pay off big.

This same method can be applied to anything really, mobile apps/games on FB, lead generation, sweeps, ecommerce and more.

When you find the unicorn (in this case lets say a shirt where the CPA is $2) you want to scale it out horizontal – more interests, more countries, more devices, more ad placement types, etc.

STEP 1 – Look at your data, see if there’s any significant pattern there.

Lets say you did a shirt that was about cats, you targeted 5 different interests (remember best practice is 1 ad per ad set, and 1 interest per adset only, using same demo)
Interest 1 – cat rescue society
Interest 2 – everything for your cats store
Interest 3 – i love cats
Interest 4 – my cat is my god
Interest 5 – people that like cats (broad)

You see that – my cat is god did crazy well, what you want to do is look what age breakdown/sex breakdown that interest had, and which ad placement.

Lets say the $2 cpa came from Mobile News Feed, Age group 45-55, Female and that interest.

STEP 2) Open audience insights and plug in this shit!!

Now click on pages liked on top, and sort by affinity. You want to take the top 10 sorted by affinity, duplicate your ad set and target one of these per.

Run the data, and see what happens. When you uncover a kick ass interest, you want to repeat this step 2, and plug both the original interest, and the new one in – this helps you narrow down your audience which is cat crazy that buys everything cat related you throw at them.
Step 3) Is your shirt in english? Make an adset using your best targeting for every single geo that speaks English. *NOTE: If you are running geo specific offers, this might not apply

Step 4) Plug in your top performing interests into audience insights, select NO demo targeting – and see where most of the fans for these pages are from. Target these geos using existing ads and/or new ads translated into their languages.
There you have it, some tested & tried methods to quickly scale. Happy New Year!!

  • What about targeting 5 demos at once and then looking at the data… can’t use see which interest had the most clicks/sales/engagements in the report ?

    • Two reasons why thats a bad idea

      1) it wont let you spend the same ammount per interest that way.. one may always drain most of the budget

      2) their reporting doesnt allow you to break down by interests when you group them

  • super

  • Mircea

    Hey Attila! Great Post! I have one question:if an interest found in audience insights using your method has only 20k potential reach is it ok to use it? or there is a minimum for this to generate good results? Thanks! @@iamattila:disqus

    • It all depends what you are promoting. 20k is tiny for shirts but could be huge for high ticket items

      • Mircea

        Thanks for this quick answer! Actually i think i didn’t ask the question very clearly …what i wanted to know is : does facebook optimize differently on a smaller audience than on a bigger one…for example: if i have 6 20k interests is it better to stack them in one ad set or just make a separate ad set for each of them? Does having stacked interests help you? thanks!! @iamattila:disqus