When to Turn Off a Facebook Ad? My 10, 20 and 30 € Testing Thresholds

Short answer: I check every Facebook test campaign at three spend points. At 10 € I look at CPC, and at 1 € or more I turn it off (09:27, 09:39). At 20 € I look at add to carts, and at zero I turn it off (09:59, 10:15). At 30 € I look for a purchase, and with no purchase I kill the ad (10:31, 10:43). Before that I warm up the new ad account. Then I launch five campaigns a day, one product per campaign (04:07), each at 50 (04:48). Every campaign is CBO (05:40) with one ad set (05:45) and three ads (05:52) of different types (07:44). I walk through the whole process in this YouTube video.

Stuck on a prompt? Post it on my free Discord server and ask. Join the Discord →

This is how I advertise, among others, a Canadian store. At the time of the video it had reached 4,000 Canadian dollars in one day (00:09, 00:12). I had launched it something like two weeks earlier (00:15). That is one store, not a typical result. This is not a promise of results. Outcomes depend on product, budget, execution, market and experience.

When to turn off a Facebook ad: three spend thresholds

I turn off a test campaign at one of three points, counted in euros spent.

Spent	What I check	When I turn it off
10 € (09:27)	CPC	CPC of 1 € or more (09:39)
20 € (09:59)	ATC, add to carts	zero add to carts (10:15)
30 € (10:31)	purchase	no purchase after 30 € (10:43)

If there is one add to cart or more at 20 €, the ad keeps running (10:21). A 1 € CPC is too much for one ad in my book (09:53).

These are general thresholds. At the first and third threshold I say plainly that it also depends on other things (09:50, 10:52). I don't have your numbers or your ads in front of me, and every ad will differ a little (08:58, 09:11).

How to warm up a new Facebook ad account

On a new ad account I never throw in five campaigns at once (03:51). I warm it up slowly, one campaign at a time.

Day	What I switch on
1	one campaign (03:32)
2	a second campaign (03:40)
3	a third plus a fourth (03:43)

I always do this on every new account (03:27). How many days the warm-up runs after day three, I don't say in the video.

How to set up test campaigns: five products, a 50 budget, CBO

After the warm-up I launch five campaigns a day, one product per campaign (04:07). Each campaign is a different product. For example a t-shirt or polo, jeans, a cap, a jacket and shoes (04:25, 04:43).

A one-product campaign sends the click to that single product page (02:40). The store itself doesn't have to be a one-product store (02:17).

I launch every campaign at 50 euros or 50 dollars (04:48, 04:52). That number never changes for me (04:59). The structure is identical every time (06:01):

• plain CBO, nothing else (05:40),
• one ad set (05:45),
• three ads in the ad set (05:52).

I always optimize for purchases (12:31). How to build the campaign structure from scratch is in my post on Meta Ads for ecommerce in 2026.

The second campaign type I run is a collection campaign. The ad sends people to a collection with, for example, 40 or 50 products (02:55, 02:58). I don't show how I test it in the video.

Can you test Facebook ads for less than 50?

It depends on the market (05:05). For bigger markets like Germany, the UK or the United States I wouldn't go below 50 (05:13). For Poland you can spend less, because ads there are much cheaper (05:16, 05:21).

How much exactly is enough for Poland, I don't calculate in the video. I run 50, so that is the only thing I talk about (05:27). What ads cost a store and what to do when Meta blocks your account is in my post on Facebook ads cost.

Which ads to test in one ad set

The three ads in the ad set must be three different types (07:44). That matters a lot to me (07:50). These are the types I use:

1. One big main image with several smaller ones next to it (06:19, 06:26).
2. A carousel you swipe through image by image (06:43).
3. A standard ad with a single product image (07:04).

I change the background too. Sometimes white, sometimes a store, sometimes a fitting room (07:19, 07:24). Different types in one ad set give me a comparison: the next day I take the type that won (19:07). The glossary covers creative testing in more detail.

What is a good CPC and CTR on Facebook

The three numbers I care about most are CPC, CTR and ROAS (07:57, 08:01, 19:51). I also look at link clicks (08:19). If there were no sales, I analyze only CPC and CTR (17:40).

Metric	My read
CPC	good below 1 € (18:01)
CTR 2-3%	average, that's good (18:06, 18:11)
CTR 4% and up	excellent (18:14)

Most of the time you will see 2-3%, and that is fine too (18:21).

Example: an ad had a 0.62 CPC and a 2.7% CTR (18:43, 18:47). It made no sale (18:51). You can still see the ad is good and worked (18:55).

Why 50 conversions: the Facebook learning phase

For me, 50 is the magic number on Facebook (11:58). My explanation: when you pick the budget, Facebook says the ad has to go through 50 conversions (12:04, 12:14). A conversion is the event you optimize for. For me it is always purchases, because I want people to buy (12:31, 12:37).

Until the ad gets through the learning phase, it keeps learning (12:58, 13:08). Only at 50 does Facebook know exactly which people to show it to (13:19, 13:29). Sales still come in during that time. Facebook just keeps picking the best people (14:09, 14:13).

Should you add new ads every day?

I used to add new ads every day (11:20, 11:24). That is how I show it in my free course on YouTube, which runs almost two hours (11:08). I don't do that anymore. I give Facebook time to optimize the ad and I watch how it learns (11:30, 11:38).

Why Facebook ad results are inconsistent

Facebook keeps learning and keeps looking for new people (15:50, 16:03). A 50 euro budget is a window of people Facebook matches your ad to (14:28, 14:34). Scale to 100 € and the window doubles (14:49, 14:55). Facebook shows the ad to twice as many people and goes through the learning phase again (15:01, 15:12).

That inconsistency is normal (16:13). I'd say you will have it until about 30 sales (16:21). Facebook pushes ads to new people who may not buy, because it wants to find out whether they buy (16:32, 16:37).

What to do with the ad that won the test

From the winning ad I build three new ads of the same type the next day (19:17, 19:20).

Stage	What I do
Test day	5 campaigns with 3 ads each, so 15 ads to analyze (20:51)
Next day	three ads only in the winner's type, each with different images (19:20, 19:37)
The day after	CPC, CTR and ROAS again, I pick the best ad format (19:51, 20:04)
Days four, five, six	new ads of the same type (20:29)

I pick the product the same way. Example: the t-shirt or polo did best (21:01). The next day, instead of five different campaigns, I run, for example, three t-shirt campaigns (21:14, 21:19). The cap came second with a very good CPC, so I add cap campaigns (21:39, 21:46). Each campaign is a different t-shirt or a different cap (21:53).

Why every ad account sells something different

At the time of the video I had almost 30 ad accounts (22:33). Different accounts sell different things. For example t-shirts on one, shoes on another (22:40, 22:48).

There is one rule: you keep feeding the algorithm more of what already works (23:21, 23:24). The same ad type and the same product type, repeated day after day (24:06, 24:15).

I recorded this during Black Friday. The way I advertise doesn't change much in that period (01:18, 01:27). Ask me about ads on Instagram or in the comments under the video, I answer everything (01:01, 01:03). We also talk ads in the community.