Hello everyone! Today we will talk about a fuss from a direct advertiser M1 and a leak through the advertising network TikTok. But first, a little background...
It's no secret that since 2022, the TikTok advertising platform has taken the optimization of advertising very seriously. Just like in Google or FB, they began to tighten the screws on the interior-offers and other advertising with black approaches.
In our team, the main focus has always been the gut, which we poured on different geos. The offer in question had already been cast successfully by us before the guys from M1 offered excellent conditions in March of this year. They were difficult to give up, so I decided to resume the tests. What came of it – read more...
Full information on the offer
- Work time: 19.03.22-13.04.22
- Offer: Keto Balance - capsules for weight loss
- Bet: $35
- Traffic source: TikTok
- GEO: Spain
- Expense: $3121
- Income: $11235
- Profit: $8114
- ROI: 259.98%
On Keitaro, you can see the full volume of leads, but, unfortunately, we do not pull the expense there. For this, the team has its own service. Everything is clearly shown on the second screen.
A couple of service parameters are covered up, as they are tax and team policy information.
Approaches to the bay
For the bay, I used an agency business center, in which I received offices. The cabinets themselves do not require farming, resting or anything else. The flooding took place in various ways. At the time of testing new creo's, they were filled manually in a small volume according to the principle: one campaign, where 10 adsets, 2 creos each. There are 3-4 such campaigns per cabinet and up to 10 cabinets per test.
Yes, yes, do not forget that our team has strict rules: 60+ Creo (minimum) per day from a buyer. The visual that showed itself well, I uploaded, duplicating each 5-7 times. After the third run, the selection was reduced to the top ones, and I uploaded these creatives to other cabinets according to the principle of tables (the functionality of the TT itself), making many copies of the best ones. And then, according to this scheme, he propagated it in the offices. Even taking into account that I was constantly at the computer, I put auto-rules for reinsurance with the conditions I needed, which TT itself also has.
Screenshots from offices
PS Excuse me here, but TikTok itself scales due to cabinets, and since they get banned relatively quickly, it’s unlikely that you can upload screenshots of all cabinets here.
Settings inside the campaign + target
Let's start with age. Here is very important! I started to pour 25-55+ and received feedback from the guys from M1 on leads, which is a lot of deviation from the audience of 25-28 years old. There were several reasons: sometimes they don’t buy it, because it’s expensive, sometimes there is no money, etc. Another important reason for failures is the lack of delivery. There are islands where there is no delivery within Spain, hence there are many deviations. This problem was quickly solved by putting them in the GEO exception. Fortunately, the TT settings allowed this (but a lot of things cannot be excluded from the TT).
After I decided to rearrange the age of 35+, in a couple of days, according to the daily dynamics, the traffic dipped. The fact is that the main audience of TT is people aged 25-35+.
After 3 days, it was decided to return the age. The deviation was from the age of 25-28, but in TT after 25 the lower threshold is 35+, which means that the entire audience from 28 to 35 years old is lost, and this is a lot. When I returned the age of 25-35+, the volume of useful leads increased noticeably, but the deviations did not go away either. However, it was worth it!
Internal setting
- Event : Payment made, card added
- Placement : TikTok
- Gender : M, F
- Age : 25-55+
- Interests : no interests
- Bid Strategy : Standard Bid
- Optimization goal : Conversion
- Budget : Any
- Automatic optimization of creatives +
Ad setup
- 2 creo,
- 1 icon (any),
- 5 descriptions,
- button to choose from : dynamic or standard with two actions.
Creative approaches
During the low tide, he did not use either a proclus or a cloaca. Although Keitaro's screenshot shows a campaign with a cesspool, it did not perform well - the traffic was cut a lot.
The very principle of the ebb was in the bleaching of the land + creo without “before and after”. I haven’t put product labels on Creo yet.
If after some time the Creos ceased to undergo moderation, then there were two solutions:
- we run it through an uniqueizer or directly in the video editing program, we render again to obtain new video metadata;
- in the video editor, directly on the video itself, we impose frames, fog, rain effects and, if they are too bright, then lower the transparency so that the creative can be seen. It also helps when converting creatives to b/w.
Difficulties encountered throughout the period
Many creatives in a short period
The best creatives are tested every day, even if you already have a working link. Scaled, but continued to make Creo.
Site selection: pour for "Complete payment" or "Add to cart"
Both events work. But “completed payment” spins traffic, consistently choosing an audience, and “add to cart” spins very quickly, giving a lot of traffic, and for it, campaign rates should be lower. And keep a close eye on it! I would even say that you need to feel it, otherwise there will be torsion, and you won’t wait for normal indicators.
Age
I wrote above that there was a lot of rejection, I had to sacrifice and pour it as is.
Many variables were tested within the TT cabinet
Which button is better, to pour only on TT or on other sites that the cabinet offers (like Pangle), standard bidding strategy or the lowest price, with or without targeting, etc.
The only difference in the buttons is that if you choose dynamic, you cannot upload the campaign to the table after uploading. It is also impossible to upload a table to a campaign with a dynamic button, since there is no such parameter in the table, and it will generate an error. And in terms of the best performance, buttons do not matter. All work equally well. If I chose static, I usually used “in store / get quote” (to the store / how much it costs). The "details" button also works fine.
Results
Even taking into account the deviations (there are many factors here: the age of the audience, delivery, etc.), the approval itself is crazy! Relatively low cost per lead, ROI is very pleased. Payout per lead of $35 with a great bump. In general, it turned out cool.
Thank you!
