The Most Active and Friendliest
Affiliate Marketing Community Online!

“AdsEmpire”/  Direct Affiliate

An extremely helpful Facebook ads case study

OscarMike

Active Member
Hi all, I recently watched a case study on Youtube about Facebook ads that I'd like to share with you

The campaign is for an ecommerce brand

It takes 3 days to become profitable.

The product is being sold for $30.
The breakeven ROAS is 1.5.

Day 1:

The average ROAS for all adsets is 0.4
The average cost per purchase: $62
There are 7 active adsets
There are 4 purchases across 3 adsets

Ad set 1: Amount spent: $120 - He got 2 sales
Ad set 2: Amount spent: $71.11 - He got 1 sale
Ad set 3: Amount spent: $40.90- He got 1 sale

The reach for all adsets is above 10,000
CPC under $1 for most adsets

Day 2:

Adset 1: amount spent - $801.49 - cost per sale: $1,135 (I'm not sure about this but this is what his screenshot showed).
32 sales. ROAS 1.42
Adset 2: amount spent: $69.42 - 2 sales - cost per sale: $69.44 - ROAS of 1
Ad set 3: amount spent: $38.84- 2 sales- cost per sale: $68.89 - 1.77 ROAS
Adset 4: amount spent: $16.14 - 2 sales - cost per sale: $69.44- ROAS 4.30
The cpc for adsets: all under $1

Day 3:

Adset 1: amount spent $1,327.58 - 67 sales - cost per result: $19.81 - ROAS 1.82
Ad set 2: amount spent $36.02 -5 sales - cost per sale: $7.20 - ROAS 5.75

As you can see, the ROAS increased above the break even ROAS of 1.5 over 3 days but it started out below the breakeven ROAS

The cost per sale decreased from being unprofitable to profitable over 3 days. This is due to Facebook optimization.

Part of running of successful marketing campaign is knowing how to define success and failure. You need to know how to define success and have the right expectations for all 3 days of the campaign.

The person who ran this campaign got 4 sales after spending about $200. if you don't get these initial purchases, you stop the campaign. There's not point in spending $2000, if you don't get any sales after spending $200.

I hope this helps.

Feel free to share your thoughts. Thanks.
 
Last edited:
The person who ran this campaign got 4 sales after spending about $200. if you don't get these initial purchases, you stop the campaign. There's not point in spending $2000, if you don't get any sales after spending $200.

I hope this helps.

Feel free to share your thoughts. Thanks.
My rule of thumb is if you land 300 - 600 referrals and do not see a few conversions with RTB/DSP traffic either the traffic is garbage or the offer sucks ??? Don't beat on a dead horse --pivot --there are exceptions ...


100 - 200 Facebook referrals should show some profit easily (for most offers maybe selling Bentley or Ferrari autos might need more :D ).
 
My rule of thumb is if you land 300 - 600 referrals and do not see a few conversions with RTB/DSP traffic either the traffic is garbage or the offer sucks ??? Don't beat on a dead horse --pivot --there are exceptions ...


100 - 200 Facebook referrals should show some profit easily (for most offers maybe selling Bentley or Ferrari autos might need more :D ).
This is extremely helpful insight. Thank you for sharing Graybeard.
 
Part of running of successful marketing campaign is knowing how to define success and failure.
I agree with it. Incremental success is important. That should be the mindset
 
Part of running of successful marketing campaign is knowing how to define success and failure.
I agree with it. Incremental success is important. That should be the mindset
Excellent point. When I first learned about Facebook ad testing, I mistakenly thought we had to spend $2000 to see 1 sale. That's not the case.

It's advisable to spend in increments first to see positive results.. eg spend $200 to see 4 sales. PPC Coach advises seeing at least 3 sales after spending $100. If not, I don't continue the campaign.
 
Last edited:
The money does not matter the numbers matter. If you get 30 referrals (clicks on PPC CPC ads) and get a median of 1.7 sales per

get 1000 clicks then you can apply 'money' values.


; (1000/30)*1.7
~56.67 Sales conversions
;

All of nothing = nothing
 
MI
Back