The Most Active and Friendliest
Affiliate Marketing Community Online!

“Adavice”/  “CPA

Question about Facebook optimization

OscarMike

Active Member
Hi all, I've read about the three 5's of Facebook marketing:

1) I need to spend at least $50 on an adset for the machine learning algorithm to properly optimize
2) The adset's audience size needs to be at least 5 million for the algorithm to optimize
3) I need at least 50 sales in 7 days for the algorithm to optimize

Is it possible to reduce a high cost per sale to the point where my campaign is profitable?

Right now, my cost per sale is $62.38 and my profit per unit sold is $13.02 (taking into account COGS, platform fees and shipping fees). I'm selling a mug for $26.95. My ROAS currently is around 0.5.

By getting 50 sales in 7 days, can I get my cost per sale to be under $13.02 and actually make a PROFIT?

Thanks
 
Spend a couple of hundred and find out?
@T J Tutor might understand better if that is workable in your ROAS situation?
Does Meta AI voodoo really work?
Thanks for the reply Graybeard. According to Gabriel St Germain, who is a facebook ads "expert" on youtube, I need 50 sales in 7 days for the algorithm to optimize. If my cost per sale right now is $62.38, I would theoretically have to spend $3,119 in order to see if I can reduce my cost per sale. According to one Facebook article, it could drop at least 40%.

Just to clarify, are you saying that I can see my cost per sale drop significantly after spending only a couple of hundred dollars instead of $3000?

Thanks
 
Last edited:
1) I need to spend at least $50 on an adset for the machine learning algorithm to properly optimize
Just to clarify, are you saying that I can see my cost per sale drop significantly after spending only a couple of hundred dollars instead of $3000?
My statement was based on your statement 1) that is up to 4 ads. I would try the best converting on to date first.

You don't say how many views to date for each ad and the conversion rate for each ad You just talk about some expert at YouTube --if he is so damn good why is he whoring on YouTube? Seems he could make more...

Update: Gabriel St-Germain
look at the most recent post :D
1706792510982.png

1706792809068.png



Do you ever check these people out that hype on YouTube?


I asked myself the same questions as they did at reddit 4 years ago.
 
\
My statement was based on your statement 1) that is up to 4 ads. I would try the best converting on to date first.

You don't say how many views to date for each ad and the conversion rate for each ad You just talk about some expert at YouTube --if he is so damn good why is he whoring on YouTube? Seems he could make more...
Hi Graybeard, the ad that has a 0.5 ROAS has 4934 impressions. My ad has 2 conversions so far with 46 landing page views. 2/46 = 4.34% conversion rate.

I've read that I need the ad to at least have 10,000 impressions and to let the ad run for 48 hours.

Do you agree with this? Thanks.
 
Last edited:
; 46/4934
~0.00932306445074989866
; or
; (46/4934)*100
~0.93230644507498986623 %

;
My ad has 2 conversions so far with 46 landing page views. 2/46 = 4.34%
That is reasonable.

Right now, my cost per sale is $62.38 and my profit per unit sold is $13.02 (taking into account COGS, platform fees and shipping fees). I'm selling a mug for $26.95.
Are you paying CPM or CPC and how much (respectively)?

; 13.02/(4934/2)*1000
~5.27766518038102959060 $5.28 CPM is a breakeven as it stands now ...
 
13.02/(4934/2)*1000
I check my numbers in the AI `machine`

User


$26.95 each sale revenue

GPT


Given that each sale generates $26.95 in revenue, we can calculate the total revenue from the 2 conversions and then determine the maximum ad cost to break even in CPM terms.
  1. Total Revenue: The total revenue from the 2 conversions is 2×$26.952×$26.95.
  2. Break-even CPM Calculation: We will use the total revenue and subtract the total COGS (already calculated as $26.04) to find the net revenue available for advertising. Then, we divide by the number of ad views (in thousands) to find the break-even CPM.
Let's calculate the total revenue and the break-even CPM.
The total revenue from the 2 conversions is $53.90. Based on this, the maximum ad cost to break even, expressed in CPM terms, is approximately $5.65. This means that to break even, your advertising cost should not exceed $5.65 per 1000 ad views.

FOOK, I'M A GURU LMAO ... I rounded up a penny :D

Flash sale Next 24 Hours!


1706799524729.png


Where does the actual problem lie? The AI Guru cannot predict. Facebook's `AI` bullshit is probably typical hype to sell more ads. "I am here from META HQ. I am here to help -only if you spend more money (sneer)."

Roll the dice --budget another CPM campaign using a reduced price Flash Sale tactic --reduce the bid 15% stop the campaign in 24 hours. Then stop and crunch the numbers.
 
This is helpful information Graybeard. Thank you.
Someone once told me that adding testimonials to my product page can improve my conversion rate.

What are your thoughts?

Thanks
 
Someone once told me that adding testimonials to my product page can improve my conversion rate.
a<=>b test with some DSP ad network traffic NOT with Facebook Ads those are too expensive for any statistically accurate volume. use a flash sale pitch on one set, c<=>d regular pricing. The look at the difference is activities.

Use Google-Analytics for those.
Scroll would be an event to watch.
Mention many satisfied customers have been real positive about our "custom coffee cup" on one version
Then try the `fake reviews pitch`

We will see, that means nothing or smells phony to me --but I am more savvy and cynical than the normal viewer ...
 
Also, PPC Coach recommends turning off an adset when the breakeven limit is reached but I have no sales. So in the case of my mug, I would turn off the adset if there are no purchases after spending $13. This way I can find a profitable adset.

The problem with this, is that I have tons of adsets that get turned off because the limit is reached. So far, I've spent over $300 and all my $13 adsets have been turned off.

Is this a wise strategy? Thanks.
 
Last edited:
a<=>b test with some DSP ad network traffic NOT with Facebook Ads those are too expensive for any statistically accurate volume. use a flash sale pitch on one set, c<=>d regular pricing. The look at the difference is activities.

Use Google-Analytics for those.
Scroll would be an event to watch.
Mention many satisfied customers have been real positive about our "custom coffee cup" on one version
Then try the `fake reviews pitch`

We will see, that means nothing or smells phony to me --but I am more savvy and cynical than the normal viewer ...
This is an interesting tactic. Thanks for sharing.
 
three 5's of Facebook marketing

Well, the 5 S's of marketing are:
Identify opportunities to sell more, serve better, save money, speak clearly, and sizzle with excitement.

However, the 5's of FB marketing are a more sophisticated, maybe not as effective, stack of goals for a funnel and is used by most of the ranking FB media buying teams. There is a fair amount of controversy regarding this in the industry that finds it limiting for some aspects and not applicable for many types of campaigns, yet it remains as the go to campaign framework for the majority of the FB media buying teams. Personally I feel it to be a dumbed down approach to enable ease of control over media buying teams until some of the team members can be singled out for more sophisticated and individualized approaches.

Then there are the variations. You see META (FB) came up with the META Performance 5 as "Best Business Techniques" from which media teams altered to fit their individual flavors of the 5.

The generally accepted FB 5 is as follows:
  • The Product
  • The Funnel
  • The Creative
  • The Destination
  • The Data
Frankly, in my opinion, there is nothing new in this. This is truly, and unequivocally, a mind numbingly simplified set of points. Anyone that has some respectable business experience already knows these points and has developed a much deeper set of points and practices.

Is it possible to reduce a high cost per sale to the point where my campaign is profitable?

I find that statement discomforting. Using the phrase "high cost of sale" indicates to me a sale structure based on incompatible elements.

Right now, my cost per sale is $62.38 and my profit per unit sold is $13.02 (taking into account COGS, platform fees and shipping fees). I'm selling a mug for $26.95. My ROAS currently is around 0.5.

This s a bit confusing.

The simple truth by which to develop suitable expectations with FB is generally 2x to 5x ROI and 2x to 4x on ROAS. You will generally see these two close hugging one another in proper funnels.

I see these "gurus" constantly changing their names, their games, their targets, etc. when the current pitches they have begin to run dry, when followers begin to see recurring funnel failures, complaints, and negative reviews.

According to Gabriel St Germain, who is a facebook ads "expert" on youtube

His last video uploaded on YouTube channel was over 4 years ago, and meanwhile, his other social media account on Facebook, Instagram isn’t active anymore. Also, all of his Shopify stores are closed, such as Flexposture.com, Cloudcushion.co, BeautyCharcoal.com.

This guy was just a churn and burn guy. He would slam the market with less than appropriate practices until he was overwhelmed with them and decided to move on into product development, whatever that means. He is the proverbial "flash in the pan" when it cam to coaching those new to eCommerce. The people I know in that segment of the industry have little to no respect for him.

Roll the dice --budget another CPM campaign using a reduced price Flash Sale tactic --reduce the bid 15% stop the campaign in 24 hours. Then stop and crunch the numbers.

This is always the case. Always go into any platform with Graybeard's "Roll the Dice" scenario, but with a real "set of brains" of FB/META tools and policies to get established on FB. I generally allot 48 to 72 hours per campaign test, but with split testing the day parting.
 
The simple truth by which to develop suitable expectations with FB is generally 2x to 5x ROI and 2x to 4x on ROAS. You will generally see these two close hugging one another in proper funnels.
with better quality ads SEM, Social Media, Premium Native Ad placement 100 to 200 views per landing page is a large enough pool to form a reliable statistic or at least a good approximation. Upgrade your creatives --that would increase your ad CTR assuming there is a good market for the product --throw some real money or some serious time on that and test them with DSP ad network traffic first.
 
Why --run with the winners and drop the losers ...
PPC Coach uses automated rules to turn off adsets that don't meet his requirements.

For instance, if the break even is $15, his rules turn off the adset if $7.5 has been spent and no add to cart. He also turns off the adset if $15 is spent and ROAS is less than 1.5, and he turns of the adset if $15 is spent and purchases are less than 1.

His strategy is to test as many possible interests as possible.

The problem I have with this strategy, is that just because an adset gets a purchase before spending $15, there is no guarantee that it will get another purchase after spending another $15, and so on.

I've only got 1 sale from this strategy.

I've had more success by spending at least $50 on an adset, instead of $15.
 
Again I go back to the law of large number is statistical analysis.

With most lower quality, and less expensive ads (on a CPM or CPC basis), I will get to 700 to 1000 (my) landing page opens, or direct offer (whichever is the case) and then cut the campaign off if the ROAS in not at least 0.33 to 0.66 --in that range, maybe the whole campaign could be optimized? Worst scenario I am out $80 or $110 and MY TIME <- My time is worth more than the ads setting all this up.

Give the campaign a chance of optimization. VERY IMPORTANT! Never beat on a dead horse --pivot and learn what you can by your failures.
The problem I have with this strategy, is that just because an adset gets a purchase before spending $15, there is no guarantee that it will get another purchase after spending another $15, and so on.
There is an old saying: It's always a sure bet that a pig will step in shit in a barn yard.
1706901248164.png


I guess it understands typonese :D

That is also how a con artist sets his mark up! ;)
 
Again I go back to the law of large number is statistical analysis.

With most lower quality, and less expensive ads (on a CPM or CPC basis), I will get to 700 to 1000 (my) landing page opens, or direct offer (whichever is the case) and then cut the campaign off if the ROAS in not at least 0.33 to 0.66 --in that range, maybe the whole campaign could be optimized? Worst scenario I am out $80 or $110 and MY TIME <- My time is worth more than the ads setting all this up.

Give the campaign a chance of optimization. VERY IMPORTANT! Never beat on a dead horse --pivot and learn what you can by your failures.

There is an old saying: It's always a sure bet that a pig will step in shit in a barn yard.


I guess it understands typonese :D

That is also how a con artist sets his mark up! ;)
This is a really smart idea. Thanks graybeard.
 
a<=>b test with some DSP ad network traffic NOT with Facebook Ads those are too expensive for any statistically accurate volume. use a flash sale pitch on one set, c<=>d regular pricing. The look at the difference is activities.

Use Google-Analytics for those.
Scroll would be an event to watch.
Mention many satisfied customers have been real positive about our "custom coffee cup" on one version
Then try the `fake reviews pitch`

We will see, that means nothing or smells phony to me --but I am more savvy and cynical than the normal viewer ...
Thanks for the response. Would you recommend split testing using a PPE campaign on Facebook? I've done a worldwide campaign for PPE- I split 3 different ads inexpensively and see which gets an add to cart, or at the very least, which one as the highest CTR. My cost per engagement is around $0.01.
 
at the very least, which one as the highest CTR. My cost per engagement is around $0.01.
That seems ok cost wise to me.
Basically, with a Facebook PPE ad, you’re paying for engagement on your posts. Engagement can take the form of video views, reactions, comments, shares, link clicks, page clicks, and assorted other clicks if you’re using a non-standard post format.
You need to pay for your intended GEO's --where you will sell and ship to.
$10 per thousand human reactions is plenty 500-600 is enough for a statistic that is meaningful.
Any event? Like, click, follow or comment ?
 
banners
Back