Advanced Meta Media Buying I:

5.7) OPTIMIZING I (Theory)an

Structure of this module:

1. Defining what optimization really is + Examples
2. Some rules and best practices before going deep
   How Meta wants you to optimize
   
3. Optimizing a Testing Campaign: ABO
4. Optimizing a Scaling Campaign: CBO
5. Optimizing a Scaling Campaign: Advantage+
   
6. Automatic Rules?

1) Defining what optimization really is + Examples

“The ongoing process of refining and adjusting advertising strategies, budgets, targeting parameters, creatives, and other campaign elements to maximize performance and achieve specific business objectives.”

Optimization is basically making sure you stay ON TRACK.

On track with business objectives. (KPIs Panel)

And that’s ALL about Decision Making.

Typical Daily Optimizations:

• Increasing or decreasing Meta ad budget
• Turning OFF loser ads / ad sets / campaigns
  
• Decreasing retargeting budget if the frequency is too high, or make the audience bigger
• If RET does not spend = Make the audience bigger
• Controlling the frequency in prospecting (make sure it stays under 1.4)
• Increasing manual biddings if your current bids are not spending

Typical Weekly Optimizations (one time per week):

• Check competitors
• Extract conclusions from the last 7 days (for example, if in the last 7 days, video was working amazing, make sure you stay alert and make more videos for the next 7 days)
• Check your page feedback score
• Check what percentage of ads you have in the learning phase (how to check it: page 8 https://docs.google.com/document/d/1_I8h8eX6M6wTtWOvldI9PVerIpExAyuffc-FSMx2n7U/edit?usp=sharing)

Typical Monthly Optimizations (one time per month):

• Make a monthly analysis
  (Waves, ROAS, creatives that worked, CVR, …)
• From the conclusions, plan the next month
• Check how much people’s salary (end of month) affected your ROAS and revenue
• Stay alert on industry trends. For example, if it’s April 30th, you want to check if your market is ready for ad angles like “Summer is coming” or “Ready for Summer?” or if it’s August 30th, check if your competitors are already having campaigns like “Back to School”.

The main thing from this introduction to optimization,

Please react fast.

If you have an AOV of €50 and you have spent €120 on a new ad set, TURN THE AD SET OFF.

If an ad set is above target ROAS after 2 or 3 days, and you still have the initial budget, INCREASE THE BUDGET.

2) Some rules and best practices before going deep

Remember, optimization is all about decision making.

Decisions based on data, analysis, trends, competitors, etc.

The first step is how to set up your ads manager columns.

>>>> Ads Manager

The second step is understanding how Meta wants you to optimize and what happens when you change the budget in the middle of the day.

Article:https://www.facebook.com/business/help/525189242169568?id=629338044106215

• What happens when you launch a new ad set in the middle of the day?
• What happens when you change the budget in the middle of the day?
• Keyword: Prorate (it means, the proportional part)

Example

Ad set A
Daily budget 50€

• What happens when you launch ad set A at 12 o’clock midday?
  It will NOT spend 50€. It will prorate and spend the proportional part of the daily budget in the remaining time of that day.
  12 midday means we still have 12 more hours in the day, so it will spend 25€.
  
• What happens when I increase the budget to 100€ at 12 o’clock midday?
  So, the day has 24h. And from 12 at night to 12 at midday it’s 12 hours.
  If the budget until 12 midday was 50€, it means the system has been able to spend 25€.
  When you increase the budget to 100€ at 12 midday, that means you have 50% of the day left, exactly 12 hours.
  100€ in daily budget in 12 hours is 50€.
  So from 12 night to 12 midday you spent 25€ and from 12 midday to 12 night you will spend 50€.
  
  75€

• And if you scale the budget at 6pm?
  Until 6pm, the daily budget has been 50€.
  So, from 12 at night to 6pm, that’s 18 hours.
  
  You have spent the proportional amount of 50€ in 18 hours.
  
  (*Sometimes Meta spends more in different hours, let’s exclude that from this simple example)
  
  Many ways to calculate this. The easiest option is, 50€/24 = 2.08€ and it’s been 18h, so 2.08€ * 18h = 37.44€
  
  You have spent 37.44€ until 6pm.
  
  And at 6pm you increase the budget from 50 to 100€.
  
  You have 6 hours remaining in the day.
  
  100€/24h = 4.16€ → 4.16€ * 6h = 25€.
  
  You will spend 25€ from 6pm to 12 at night.
  

Do you need to calculate this all the time? No. Just understand that when you do a budget change, you will spend the proportional amount of that budget in the remaining hours of that day.

Myth: if I launch a new ad set at 8pm then Facebook will blow up all the budget in the 4 hours remaining in that day. No.

Another thing about Meta,

They care a lot about the learning phase.

The theory says, if you increase the daily budget more than 20%, you will re-enter the learning phase.

Article:https://www.facebook.com/business/help/316478108955072?id=561906377587030

List of significant edits:

• Any change to targeting (audience)
• Any change to the ad creative
• Any change to your optimisation event
• Adding a new ad to your ad set
• Pausing your ad set for seven days or longer
• Changing bid strategy

List of edits that may or may not be significant, depending on the magnitude of the change:

• Your daily budget
  (changing from 30€ per day to 35€ per day won’t re-enter the learning phase, but from 30 to 70 “will do” - 30-300€)
• Your bid
  (changing a cost cap from 20 to 25 won’t re-enter the learning phase but 20 to 50 yes)
  

*This seems to be changing. I will keep you updated, but in terms of daily budget changes, the machine is tolerating bigger changes.

_________________________________________________

Should you look at “today” to make a decision and optimize?

In 90% of the cases, no.

To keep it short, a good time frame is looking at the last 2 or 3 days. It used to be 7 days, but in my opinion and experience, the machine is now smarter, so you can make decisions in 2-3 days.

Exception: If you’re spending >1,000€ per day, AND the morning hours are usually great for you on Shopify, then you can make optimizations in the middle of the day, for example, if you see that today’s performance is bad by 2pm.

Exception: you see good tracking (not a big difference between ads manager purchases and shopify purchases) and you sell an impulse-buy product (for example, AOV lower than 20€).

But if you sell for more than 50€ and the users take on average 3 days to make a purchasing decision, why on earth would you make a lot of decisions on “today”?

For how long should you leave a new ad set active? Which means, how long does it take to stabilize?

Meta is machine learning.

Which means, for the performance to be stable, you need to accumulate a lot of purchases per ad set (50 or more, per week).

As a general rule, I leave a new ad set active for 3 days before making any decision. Or until it has spent the price of the product I’m selling.

To make a decision from a test, 3 days.

But more days if I look for stability.

_________________________________________________

3) Optimizing a Testing Campaign: ABO

• Mostly testing
• Optimizations: turn off, keep, scale
  Turn off if I have spent my AOV and ROAS < target ROAS
  Keep if it looks not bad not good
  Scale if ROAS is target ROAS or more
  
• Actions and conclusions: if it’s a winning ad, make new variations and start scaling the winning ad horizontally

4) Optimizing a Scaling Campaign: CBO

• Mostly scaling
• Before campaign turn off, correct ads
  CBO, 3 ad sets, 10 ads
  “The budget went to the wrong ad”
  “But I have an ad that spent 50€ but ROAS is amazing”
  
• Then if it’s at 200€ per day, don’t turn off, decrease to 120 and give 2 days, if it corrects nice, if not decrease to 60 for example
  

5) Optimizing a Scaling Campaign: Advantage+

• Mostly scaling
• Before campaign turn off, correct ads
• Then if it’s at 200 per day, don’t turn off, decrease to 120 and give 2 days, if it corrects nice, if not decrease to 60 for example

My opinion on Meta automatic rules, for optimization.

Stop loss only.

Avoid running a bad ad for too long and wasting money.

Make the rule based on your AOV.

Example - “If I spend my AOV and I have no sales in a 2-day period, then turn the ad set off”.

My recommendation:

2 days because if you have a CBO at 110€ in daily budget, and for the last 7 days it works amazing, but surprisingly today it didn’t get any sale, if we don’t put “2 days” then it would be paused…

So, just put 2 days to have more margin.