Home > Facebook & Instagram >

---

*The Ultimate Facebook Newbie Tutorial 2018/19 - Part 8 - Test, kill & optimize (26)

---

@stickupkid, thank you for this excellent tutorial.. just to confirm, have you ever tried setting up your campaigns in the way described in this thread - https://stmforum.com/forum/showthrea...LES-(In-Depth).. i.e. creating a separate campaign for each ad and then duplicating ad sets 1-3 times and also the ad 1-3 times.. if so, how did it work out?

---

I have tried several different setups, for some reason I stick with 2-3 ads per campaign - 1 ad set. Then duplicating or duplicating and only change images (mainly in smaller geo’s/audiences where ad refreshments are often needed)




Sent from my iPhone using STM Forums mobile app

---

Thanks for the reply @stickupkid. When you say 2-3 ads per campaign, all those ads are the same ad, correct?

---

Just different images!


Sent from my iPhone using STM Forums mobile app

---

Thanks for clarifying @stickupkid! Do you normally duplicate the campaign then one or 2 times so you end up testing the exact same ads/images 2 or 3 times (similar to what mation does)?

---

That’s correct. Running AU for several weeks now with only 3 different images, duplicating over and over.


Sent from my iPhone using STM Forums mobile app

---

When you say duplicating, is that within the same ad account? Or taking the campaign and running it in a separate ad account?

And if it is just duplicating the adset within the same ad account, do you keep the budget the same for the duplicated ad set or increase it?

---

He was referring to duplicating within same account yes. Sometimes keep budget the same, sometimes double it, sometimes even lower budget from start.

But duplicating on another ad account is another great way!


Sent from my iPhone using STM Forums mobile app

---

@stickupkid, is there any sort of thumb rule for how many times you normally duplicate a campaign with the same ad/image when first testing an offer? In addition, with an audience size of say 500,000 how many images do you think would be best to avoid over saturation? Thank you again for your help!

---

Thanks for this post @stickupkid .

You say you do not work with pixels, why is this?

---

I have done it like this for like 5 years without a pixel or cloaking.

Since 2 months we run some campaigns with pixels. CR wise not a significant difference, BUT it seems the cpc’s to be more stable for a longer period.


Sent from my iPhone using STM Forums mobile app

---

I am still a little unsure on what the optimal structure should be, and the logic behind it. From what I understood, the idea behind running the same ad in triplicate was to see the results from the ad behind shown to different pools of users, this makes sense to me. But if you have a different image in each ad, how can you know if the difference in performance was down to the pool or the image?

Let's say I go with three of the exact same ad in an adset. When I test slight variations of the same ad (e.g. changing only one of image, headline, body copy), do these go in the same adset? Or get moved to a new one? Does the new variant of the ad also get run in triplicate?

After iterating through this a few times, you'll have probably changed the image, headline, and body copy - is this now an entirely new ad, deserving of its own adset? Or, presuming that you kill the ads that aren't profitable, then I assume the adset optimisation moves along with the highest performing ad. But what if the three previous iterations of your ad are actually profitable, just not as much as your best ad, should these ads be run in their own adset to avoid competing optimisation criteria (maybe the earlier iteration appealed to a different subset of the audience)?

---

If you run 1 campaign, 1 ad set, 3 ads (all same settings, only different image) -> image 1 does bad, image 2 does bad, image 3 does well. All 3 ads were shown in a same pool, so assuming image 3 is best. Cut the expensive ones and monitor how it effects CR/volumes.

Running that winning ad in new campaigns/ad sets will confirm it's a winner (or not ofcourse). Sometimes a sure hit turns out into a miss and vice versa. So many factors influence the performance of ad... timing, competition, luck etc..

---

That's a great post, thanks for sharing.

But how do you decide when to cut an add?

Is it always your gut feeling, or do you use any specific formulas for your campaigns, like cut all ads below 4% CTR?

---

I will not cut an ad below 4% CTR if it’s still a profitable one ofcourse. Also depends on the geo/angle a bit.

My rule is to reach 100% roi at least on a campaign as a whole and in between I cut “expensive” ads and test new ones to dive under the lowest to increase ROI.
Most of time I cut ads above 0.20 ct when running a cpa offer between 6/10 euro.

Just two rules I kinda stick to then:
- at least 100% roi
- ads above 0,20ct = kill

---

Thanks for sharing the numbers.

But just to confirm, you meant you cut everything above $0.20 CPC, right?

---

Corrrecto! Unless it's a 25ct ad and my epc is 2,50 for example, enough margin there but that doesn't happen very often :-)

---

Thanks for the answer again stickupkid

---

Hi @stickupkid. Sorry being off topic.

Is it still possibble to run sweeps campaigns without cloaking and tracker with the basic survey style lander these days? The last time i tried, fb disabled my account after 3 days. Im also using non aggressive ads.

---

We still don't cloak, nor use a tracker indeedo. The FB account game is a tricky one. Nowadays it's harder to keep accounts in the air than launching a profitable campaign.

---

Yeah its hard to keep accounts alive these days with sweeps. I always try to minimize the amount of negative feedback as much as possible but still getting bans after getting a considerable amount of traffic

How long does accounts last for you gyuys in average?

---

Something like this I think on farmed ones.... (so after atleast 6 months of natural behaviour and little wh spend)

10% - less than 7 days
20% - 7 days
40% - 50 days
20% - 100 days
10% - 180 days or more

On accounts from friends/family it's different.

---

Thats interesting. Hopefully i can achieve that same statistics.
Thanks for sharing.

---

This is a REALLY interesting approach @stickupkid - I know it's been a while since this post has been replied to, but super-curious about it.

When you say that you don't work with pixels or other tracking stuff I'm curious about that...

Does that mean that you don't have pixels to track conversions and optimize by them? Not quite understanding that, because otherwise how to do tell which ad is converting if not using a tracker and/or pixels?

Are you saying you do/don't run conversions campaigns (or do you run Traffic/click campaigns instead) and don't use a pixel of any kind to optimize in a conversions campaign (if you use them).

What level of cost do you shoot for before killing your low % CTR ads?

---

Well this method works for sweeps, but I can't say for other niches. If you target LIDL likers 35+ in a certain geo, and you launch a LIDL offer you are quite on point already. When using a pixel you will see, most likely, the best performing offers have lowest cpc, just like Traffic/website. Most likely most conversion will come from the low cpc/hight CTR ads, there is no pixel needed to tell you that.

At the moment I am doing 50/50 - so half with and half without pixel. I barely check CTR, I check whats coming in and what goes out (cost/revenue) and I kill expensive ads. What is expensive? depends on your ROI goals ofcourse.

---

Thx @stickupkid for these interesting numbers.
As a newbie I'm still looking at what traffic source I would test.
Very tempted to go for push or native, because it is a new toy for me (shiny object syndrome much)

But reading this manual made me realize I know all this FB stuff basically.
Did this all for agencies (not the affiliate performance campaigns, but more branding and other vague stuff). So I would be an idiot not to start with FB because the learning curve will be much easier for me.
Biggest challenge wil be tracking, not the most technical guy in the room. I can manage some technical stuff and code... but nothing too crazy.
So hoping to find a dummyproof solution there.

Once again we Flemish/Belgians can learn something from the Dutch because thy move a bit faster

Thx

---

Home > Facebook & Instagram >