-
Why Your ROAS Goes Down When You’re Scaling Your Ads
Hey everyone,
Today I want to speak about a topic that took me an extremely long time understanding and prevented me from ever scaling up my ads.
*Before we start:*
*Every ad account is different and every ad account should be treated differently. The basic principle I’m trying to explain in this post might not be the reason why you’re facing a decline in ROAS. Many more factors come into play.*
So, let’s get started:
Why does your ROAS decrease when you’re scaling up?
And how can we combat that?
To answer the 2nd question, we first need to understand why your ROAS decreases when you scale up:
**In short: Facebook only has a finite amount of people that can convert at or below your target CPA in a given day.**
Example:
Let’s say you’re targeting a 1% Look A Like of purchasers with a video ad, and your target CPA is $20. Your daily budget is $100.
This means to stay at or below your target CPA, Facebook has to get you at least 5 Conversions every day.
5 Conversions @$20 per day looks pretty reasonable (Given that you have a great ad, proven offer, and great conversion rate)
Let’s assume you’d scale this Ad from $100/d to $1000/d because you are happy with your $20 CPA.
It is much harder for Facebook to get you 50 Conversions a day for $20 per Conversion.
Why?
**There are only a finite number of people on Facebook/Instagram each day.**
And these finite number of people see hundreds of other advertisers out there who “battle” for the attention of the people you want to convert.
On top of that comes, that not every person in your audience is a potential customer or is even ready to buy yet.
So, back to our 1% Look A Like Example with the Video Ad:
From your 1% Look A Like, Facebook may think only 100,000 people are worth advertising to, because Facebook thinks the rest is not a good fit for your ad/product
*Note: 100,000 is just an estimation for this example. However, it is true that Facebook will only target only a subset of your audience, even if you run a 1% Look A Like Audience*
From those 100,000 people, only 1,000 may be potential customers, and from those 1,000, only 100 might be interesting in buying something today. And these 100 people also get to see other ads.
*(Facebook has a very clear understanding of how frequently people buy, at what time they buy, what they buy, how long ago they last bought something, and much more)*
A much more realistic CPA Example then may look like this:
Facebook can get;
11 Conversions @$14
10 Conversions @$20
8 Conversions@$23
6 Conversions @$27
10 Conversions @$30
This equals $1000 per day. However, now we have a problem.
Our average CPA now is: $22.22
So, now our CPA is above your target CPA.
Now that we’ve understood why our ROAS *usually* declines when you scale up, we can begin to answer the 2nd question:
**How can we now reach a lower CPA again at a larger scale?**
​
1. **Create different types of creatives**
Not every person reacts to the same type of creative. A few examples:
Some people buy more from image ads than from video ads. Some people like close up product shots, some people like emotional UGC videos, and there’s much more.
Let’s say you now introduce a great image ad to the video ad from our example. This image ad reaches a different subset of people in the 1% Look A Like
*(No doubt there would be an overlap with the video, but all in all, it will reach different people)*
So, our situation would look like this:
Audience: 1% Look A Like
Ads:
Video Ad with $500/d budget
Image Ad with $500/d budget
Given the example from above, Facebook now could get
11 Conversions @$14
10 Conversions @$20
6 Conversions@$23
because it doesn’t have to spend all of the $1000 just on the video alone.
This equals $492 a day, and our CPA would now be:
$18.22
This means now our video ad again would spend $492 at a CPA of $18.22
For the purpose of this example, let’s assume the new image ad can get the same amount of conversions:
11 Conversions @$14
10 Conversions @$20
6 Conversions@$23
Our image ad is now also able to spend $492 at a CPA of $18.22
This now allows us to spend $984 every day at a CPA of $18.22. For the $16 we have left, let’s say the image can get another cheap conversion @$16.
This would then give us an overall CPA $18.18 for our $1000/d budget.
**2. Introduce new audiences**
Another way to combat declining ROAS is to introduce new audiences. You might want to try 2% Look A Like Audiences, Bigger Interests, or just Broad Targeting (Just age and gender set)
​
Different/Bigger Audience = Different/More People To Reach = Not only 1 Audience has do the heavy lifting = Cheaper CPAs overall.
​
These 2 strategies, are in my opinion, the easiest way to combat declining ROAS when you’re scaling up. Though, there are many more. But going over them in this post would take too long.
​
**TL;DR: Facebook only has a finite amount of people that can convert at or below your target CPA. This is why it is crucial to introduce new audiences, creatives, ad copies, headlines, and more when you plan on scaling because these reach different subsets of people.**
​
*DISCLAIMER: Every ad account is different and these are by no means all the reasons why you might be facing a low ROAS. Though, I hope this post could help you understand better how Facebook Ads work.*
​
If you found this post helpful I’d appreciate an upvote. If you have any questions, feel free to leave a comment below.
-
8 Replies
-
For anyone who knows economics, this can all be summarized by the [law of diminishing marginal returns](https://en.wikipedia.org/wiki/Diminishing_returns).
The advice here is good, as individual creatives and audiences also have diminishing returns, so when increasing spend you should increase your scope to minimize these marginal losses. That said, you still shouldn’t expect to see equally strong results when scaling simply by adding more creative or expanding audiences (unless your original creatives/audiences were poor). In almost every scenario, each additional dollar spent on advertising will be less efficient than the previous dollar.
-
I like those perspectives but for truly large businesses (>XXXm €$) „there is only a certain amount of good traffic“. ??♂️
-
What about when you’re not scaling but a good performing ad set/audience stops converting after a week or two of good ROAS? Should you just introduce new creative too and relaunch it?
​
edit: This is for smaller individual budget ad sets anywhere from $30-60/day
-
This post is a goldmine. Thanks op and commenters.
-
LOVE It!
-
Very well explained. Something that can be hard to understand if it’s new to you.
-
Excellent post, OP.
For anyone who enjoyed that post, I recommend your next read should be this blog post by Mack Grenfell: [https://mackgrenfell.com/blog/mathematical-marketing-one-piece-of-calculus-that-can-change-the-way-you-advertise](https://mackgrenfell.com/blog/mathematical-marketing-one-piece-of-calculus-that-can-change-the-way-you-advertise)
As the OP outlined above, Facebook (or any other marketing service) will generally return lower ROAS the further you try to scale, due to the finite number of people available to advertise to who can be bought within your target CPA.
As well as knowing how to bring your CPA back down again, it’s equally important to know what your optimal CPA actually ***is***.
Just as a high CPA eats up your profit margin, so too does a low CPA eat up your volume, and both are equally as destructive. Mack’s post does an excellent (if a little math-y) job of explaining how to find that tipping point and maximize your profits.
-
Great post OP. However I’m wondering, if you have 2 different creatives inside an adset, does Facebook deliver your 2nd ad to a different audience ? What I mean by that is : does Facebook try to find the ideal audience that will react to each specific ad inside your adset ? Or do you need to have a separate adset in order to make the IA change the audience (therefore it would make sense to group ads by type or different messaging in different adset to produce maximum results).
Log in to reply.