Кейс перформанс макс (чоловічі аксесуари) 2023 Performance Max налаштування
Hello everyone. My name is Yana Lyashenko and I am a Google logistics specialist. And I deliver to the target audience of the business with the necessary parameters. This means that I know how to scale your online sales due to the fact that I know how to find your target audience using Google advertising tools.
Briefly about what we want to show by analyzing this case
Today we will also look at one of the interesting cases that our team had. Why is he interesting? Firstly, I want to show it as an example… Well, perhaps this will be of interest to specialists or businesses that independently maintain their advertising accounts.
Basically, what happens when you really like duplicate feeds. Once. When you don’t have MPN, GTIN there. Duplicates. You have a wide product range. The second is if you like to make changes while the team is working in the advertising account. Two. Third, in principle, how Performance can positively influence other campaigns.
What indicators did the client come with?
What do I want to show here? Conventionally, a client came to us with a request for Performance Max. He brought very few conversions to this. Very few conversions. We started somewhere, somewhere around June 7th. Let’s take a slightly longer period so that the conversion volumes can be seen. That is, it was 35, 21, 20 sales per day.
Well, the cost per conversion was really cheap somewhere. Somewhere there is not bad, but conditionally, that is, visible. The challenge was to scale. This was literally the beginning of our buildup, then in the second month we had such a leap. We will proceed further in order.
Complexity and tasks that were at the beginning of working with the client
What was our difficulty here? The fact is that we had smart shopping running at the same time. At that moment, smart shopping was still working. Here you can already see the moment of eliminating smart shopping very significantly.
Our main task, as specialists, was not to just do a great job and leave, like we made a bunch of conversions for a date, but preferably not to break an advertising tool that already works quite well for the client.
I have loved smart shopping since the past, when it was still working. I don’t understand why Google canceled smart shopping and switched everyone to Performance. I think that this is some kind of throwback to a previous age. Because even at those moments when Performance Max was just appearing in advertising accounts and our main emphasis was on smart shopping, Performance did not suit me at all in terms of the result. But there is what we have.
In most advertising accounts, smart shopping was automatically turned into Performance at some point. But there was still a small part of advertising accounts where smart shopping continued to work. While he was there, I wrote somewhere on my telegram channel, until one day he felt sorry for them. Just quietly, silently, without even creating a replacement for Performance.
Comparison of smart shopping and performance results
That is, the task was not to break this smart shopping. Because Performance could, in principle, eat up this advertising budget very actively. Very simple. One or two transfers and smart shopping would break down. Our task was to preserve what works well enough. And this is, in principle, an adequate approach of any adequate specialist.
Secondly, the task was to try to grow from performances, without breaking smart shopping, to the same turnover in at least two or three months. Well, for example, to this one. Let’s do it this way. Here is May 15, let it be, for example, until September 10. Did Performance make how many valuables here? Well, here he made 6200 values, so you understand, smart shopping made 10 times more. 10 times more.
That is, this is a very cool advertising tool, despite the fact that the profitability is several times higher than it was in performances. Here even the performances, here even it is greater than the target profitability that you see. They are higher than what the client asked for, even at all. But still, smart shopping worked much better.
Despite the fact that there was very little need to do with smart shopping. Yes, there was a large budget on it, which, in principle, prioritized it in the issuance. But none of our performances were able to grow quickly into the same budgets. Now I will explain everything. This will be a very colorful example.
Why did you work on a duplicate feed?
That is, when you like to play with duplicate feeds, because this is more of an emphasis for those businesses that like to duplicate their feeds. We worked on a duplicate feed because smartshopping worked with the native feed. We didn’t want to change anything in the old smart shopping, whether the client or us, for example. Because we could, in principle, squeeze the client out of this, it’s right that we use the native smartshopping feed, albeit a new one.
We worked on a duplicate feed. This gave us a certain opportunity, at least some variability in our work. But still these are not native IDs. Now I will explain what this means and how it affects the work of performances. And even with numbers I will show some correlations between the dependences that occurred in this, where we even played along with smartshopping in some version. But since we play for the same team, in this advertising account, everything is only a plus.
That is, the concept of what we had is clear in principle. So, if you play with duplicate feeds, you must clearly understand why you are doing this. I use duplicate feeds in my work, but only in cases where I need to play with a certain small set of products in order to make them a cash cow in individual performances.
Or, for example, test completely different cardinal signals in order to understand what works for me and what doesn’t for the same product range. And I usually do this within the limits of some very limited product range.
Why is duplicating a feed not a good idea?
That is, full eyes that were distributed by certain experts on the Internet: “Let’s make duplicate feeds for an ordinary product one. For smart shopping, the second one is for Performance. If you have more than half of the products… There are so many consultations. performed audits.
Well, classic analytics for all projects is from 50 to 70% of the clients’ product range, products that receive less than 200 impressions per month. 200 impressions divided by 30 days. How much is this? That’s six shows a day. What sales, volumes, turnover do you expect? If your product just barely shows up six times, six times a day. Therefore, I personally don’t welcome duplicate feeds for different campaigns, for different types of advertising campaigns.
This or you do it pointwise with a specific goal, task. That is, a test of signals of any specific structure or betting strategy. I’ll understand that. If you are doing this like, well, because someone said that it will increase sales, this advertising account will now show that nothing is done that way.
Why? Now I’ll explain. This is very easy to understand. Look at what all the coolness of this Performance Max is based on, before this smart shopping. Because it uses machine learning and artificial intelligence is now being integrated very actively. Although they promised that this artificial intelligence is still used in the newest versions of Performance. But it is clear, based on the fact that the coals began to post on Twitter, that it is just about to begin to be used. Well, not very important.
What’s so cool about Performance and smart shopping?
Look at the model. In order to predict what audience to look for, plus what products to sell to people… What was the cool thing about Performance with smart shopping? He can slip your product item not only higher, but closer there or in a visible place to the person who was on your site. It’s all bullshit. Any remarketing can handle this.
And the cool thing was that he could slip this product item to a person who was looking for similar products or even analogue, analogue ones on other sites, without going to your site. Cool? This, by the way, is such a thing that when you google work clothes, sometimes quite revealing outfits pop up there from the alisa.ua store, well, women’s revealing sex clothes for special work clothes.
This is not because something was configured incorrectly in alisa.ua. And this is because somewhere there was some kind of leap in the analytics of understanding that this product can be slipped in and people will buy it. Well, it’s user behavior. It’s not bad. This is cool. Your Performance can do this too, along with smart shopping. That is, they even add an analog product item to any product range.
About the need to correctly transmit the product ID
Well, for example, I observed this in many projects over different periods of time. Swing around in this. But there is one aspect. This is provided that there is a clear connection. There is a FID. It contains the product ID. This product ID is tracked on the website. This is something we didn’t have here, and I’ll now show you how this make effect produced a certain result in smart shopping. Plus, in the performance they took a quick breath. Now I will show everything on the graphs. Well, I won’t try, I won’t try to deceive you too much with some graphs.
But the one who likes my analyzes is how to break it all down, because this case, well, this project, I don’t know, it can be broken down into 10 cases in different topics, niches. And it’s easy to find insights in each of them. This is actually a very cool project. I really liked that this idea itself, the concept of the product, in general the account itself, in principle. Because it is very radically different from others.
It is he who very clearly shows the mechanics of when some things do not match and when they need to be corrected. That is, we realized that there is a product id and there is an id of the position that it sees on the site. How it works? For example, more than once… You know that in most cases I work with foreign stores, European or American. This is very clearly visible there, especially, for example, on Shopify websites, where it is possible to transmit product IDs in different formats. Tracking on the site.
God, I’m about to deceive you. But in any case, I’ll try to leave these, perhaps, time codes. Because there are a lot of moments that I want to share. Perhaps for some this is a very scary thing, I will tell you abracadabra. But be patient.
Those who are in the know, this will be interesting for them. Here on Shopify it is possible to create these IDs in different formats. Somewhere you will see Shopify, there is underscore, USA, underscore, ID. Somewhere there’s just an ID screw. Somewhere the alphanumeric ID is transmitted. There’s SSH in there somewhere. In short, there is a wildly large number of programs from these ID guys. And it is on these projects that it is very noticeable when you generate traffic through performance or smart shopping, and then at some point the performance flounders and sharply drops in conversions.
How does performance adapt and learn?
And you don’t understand what’s going wrong there. You begin to analyze and see that even the mechanics of doing things in a performance change radically. I had an analysis of several projects where even the audience changed dramatically in two months for the performance. Why did this happen? Because look, this stupid artificial intelligence, it works on what you give.
For example, we have, for example, selling I don’t know some men’s watches there. In our case, we are generating traffic to a website. That is, it accepts this request and displays advertising based on it. There are clicks that go to the site.
If there is a performance, for example, they don’t see what comes next, for example, Add to cart. People add a cart but don’t order. He will change his targeting. If he sees that he is generating these View item, then view the product cards. But they don’t even turn into Add to cart, something is definitely wrong, which means this traffic is completely wrong. That’s the cool thing about this performance, it constantly adapts and learns.
Well, I think this is a very cool advertising tool. I don’t know what Facebook is doing there. Facebook is also praised. But this is very cool for online stores. That is, you have a living tool inside your advertising account that is constantly looking for something. And if he sees that he is not achieving results, he changes the direction of the river.
This, by the way, was discovered more than once on different projects at this point in the audience change. As soon as there is no corresponding feedback in the events. In addition to this, we have information. It is important for us to convey the performances that fill those events. So what? What does the user look at, where does he come, what does he do next with this product item? He goes for a walk on other goods or buys it or refuses. This is very relevant information specifically for automation and artificial intelligence.
Especially this performance. I don’t know, maybe there’s some Indian on Google right now who wrote this entire tool worse than me, which I admire so much. It’s a very primitive instrument. I am telling you how I understand how it works based on simple analysis and observation. And what works for me is when I see that I need to scale something up or, for example, optimize it for the better. This is what I observe.
Negative consequences of product ID mismatch
So, look, if our ID numbers do not match, then it essentially seems to generate traffic to the product card. He’s running a proper campaign. But another product is essentially like for Google, he came in, brought a person, and he didn’t buy this product item, but went through the catalog and bought a completely different product item. It’s very difficult for him to scale this up. Why? Because he doesn’t understand which product item to base on. Because he did not find this ID-book purchased in his structure. That is, the conversion is included here.
Well, for example, what I mean is that in some of your campaigns you can observe such a nuance when you have more conversions in units than at the product level. At the product level there may be much less. For example, there are 20 here, but at the product level there are only five. This may be due to… But the products and conversions are real, they are physically recorded. They are all these 20 or 10 of all conversions.
This happens because they can move on to one product item, rummaged around on your site, and bought a completely different one. I correctly assigned the Google Ads conversion to myself. He correctly stuck it into the conversion column at the campaign level. But not at the product level, because well, there was no, there was no click on this product item. She didn’t show up at all. This is very common. Especially in the themes of everything for home or decor, furniture or something like that. This is a very common story.
There is a lot of need to tinker with this structure, but conditionally. That is, for him, when you play with feeds that are duplicates of different IDs, it’s a story when it’s as if something was purchased there. Yes, this is a Google conversion, but how to use it. Where will he be moved? He does not see this product in his structure in the product group. Yes, he can partially use the information something similar.
He can slip in a similar product item, but this does not hit the mark. You see, a ten is when everything came together beautifully for us. Smart shopping used everything it had and even took advantage of the traffic generated by Performance.
Why can a performance “knock”?
But in performances the same thing happens as at Prom. Now, unfortunately, due to the fact that they do not coincide, well, at Prom the situation is a little different, but conditionally. It generates and generates traffic and then begins to sag and fall. If, in principle, everything else is ok, then pay attention to whether there are any useful actions from these products. Well, there are conversions and they coincide with product conversions.
Because perhaps your Performance puts some sales here in the conversion column, because it really was this sale from its traffic, but it cannot be compared with products. And therefore it can knock over time, because it does not have information about which one to operate. Because you have to understand, this is a very cool unique instrument, this Performance. No wonder they praise him. Not only that, several performances can emerge into symbiosis, synergy.
It happened that we excluded one experiment because we left the best one. And the one that was better fell a little by 30 percent. They can build synergy in goods. They can slip analog products. They may even wedge themselves into an auction for a completely different product line because they know that this person is about to buy your product.
This is really a very cool tool. Yes, it doesn’t suit everyone. Yes, not everyone there has super cool sales. Yes, not everyone there immediately becomes rich, as my client, a potential client, said: “Make me rich.” Not everyone. But it’s a very cool tool.
It is not possible to unify and standardize it very well in all niches so that it works for everyone. But the fact that Google tells you that this is a very cool tool is definitely yes. So that you understand the concept, this is such a big preview of what we will consider next. Because, most likely, I will break this project down into several parts that are so interesting that I want to tell you with such a twinkle in my eyes. I will teach them one by one.
initial account settings and information about the sale sold
So, in short, we had the initial data, a duplicate of a feed, a wide product assortment, a male, women’s watches there, sports hours, tactical, children’s. What did we have there? Smarts, fitness. In short, there was a lot of everything. Secondly, this is that we did not have native ID shniks, I do not know if I said. Third, do not break smartshoping. Because it was very easy to detect traffic from him. Even in some periods it was, it seems to me. Somewhere there, a slightly sagged smartshoping, and then it leveled out as soon as we quenched the appetites of existing performances a little.
What else? And of course, a client who loves makes changes. It was also some kind of contact in our advertising account. There were some periods here, because someone really loved to click at their discretion in an advertising account.
And we have a test hypothesis that has time for optimization. And you are already making further actions on the basis of the fact that you understand that a set of things was made before this. If you do not know that someone has already introduced there, then it knocks down the entire algorithm a little. You completely change the test, but in essence you need to spend it again.
What to look for when grouping campaigns?
So, what is interesting here can be gleaned? The first thing to remember is always with the initial grouping of your campaigns, try to understand who your main portrait is when buying these goods. It would seem that female and male can be grouped together. And they will somehow work there, but our task was to make this result scale.
Because the performance as soon as I started to scald at the expense of the budget, adding (!!!!!), optimizing bets, the same profitability, it began to fade away. What does it mean to fade?
Now I will show one of the graphs. Somehow here, apparently, it will be very visible. There is something here, some kind of movement is visible according to conversions. Then less, less and boiled directly to zero. This is just a schedule for argumentation, here the client was painted in a certain way. There will be all the lock, smeared, conditionally. But so that you understand what it was fading away. That is, he even stopped going into shows. There are no shows, but this general dynamics was observed.
or for example, for example, one of the brands of advertising campaigns that had to be displayed to shows in order to focus on it as much as possible. Yes, this brand was somewhat well in common campaigns. But still it was necessary to increase the number of sales on it. And here each of the performances starts. Flaws. Starts. Flaws. Starts. Fetting.
This would even be noticeable, for example, even on the basis of the same show. That is, here you can see, for example … let’s take the conversion. Let’s choose clicks. And here it is visible. Well, here is not very visible. She goes out. This Performance Max is dying.
Well, this is the period when the client will accompany it independently. This yellow. Well, there is this type of advertising campaigns, it is constantly conditionally in some depression. That is, the performance only sways and it goes out. One of the reasons was just the fact that we did not have native ID-shniks, unfortunately. Unfortunately. And now I will show how it influenced smartwatch.
Comparison of the results of smartshoping and Performance Max
I will find now. So, for example, we take such a period, June 5 – September 10, with the previous period. We have a set of advertising campaigns. They will all be smeared, but conditionally this first is a smartshoping that worked. All this is different – this is a bunch of performances that were tested at this time. Here is the brand, which I showed before that, which was included in different categories of goods or was allocated into separate branded campaigns. There were several tests there.
That is, during this period there were only 749 sales. 749. 443 of which made a smartshoping. Well, that is, the rest, 300 pieces, this was done by performances.
Not a performance, but a performance, conditionally, to understand a common common picture. Smartsoping, he is only a little by 16%, somewhere there was a little sagging. But, conditionally, it can be seen here, how different the conversion coefficient in different types of these campaigns that launched here. They all worked completely differently. Which of them is better, which of them is worse. I think that this is something under -tested, conditionally. Could be improved.
Comparison of the results of smartshoping and Performance Max by categories of goods
Or, for example, let’s see, for example, such categories. Well, for example, there is a breakdown by types of goods. And here is the company. This is all smartshoping, this worked period. It’s going further and the performances went. Maybe somewhere else smartshoping, it intersects somewhere. And so, conditionally, what we see, for example, male male times and men’s watches. That is, two versions worked here, language versions of these hours.
Well, what can I say here? Male times sank a little. Yes, then the men’s watches sank. If you look at men’s times, for example, male, male, then they are not particularly. This category. There are just a lot of these men’s watches here. Here, 200, 200, 70 are gathered. They are not too active there and stood out there. Despite the fact that the price for the conversion was quite significant in the same performances.
The same, for example, smart watches. Here smart hours, on the contrary, have grown in conversions here. Although in the performances our smart watches went very poorly. How much traffic did not generate, they went very well. Yes, this is also a smart watch. They have grown here.
Women’s clock has also grown. Well, here is a heading without category. OK. Even in order not to deceive, conditionally, you analyst, so as not to stretch it to a frantic amount. Here are male, male times. Well, male times the requests themselves began to sag. You understand why. Ukrainian -speaking well, relatively slightly grew. Smart watches are unequivocal growth. It was without categories that all performances were. Well, there is a whole porridge, most likely men’s watches.
that is, if it is visible here at 709, these are all performances. Under this dash, these are all performances. And in most cases there were men’s watches. There was a different breakdown. Tactical, inaccessible, male, brands, price tags and so on. Well, even if you look like that, men’s watches. Yes, it draws, for example, 709. Yes, but here without categories. Plus 1155.
Well, there is, well, here it shows such a high growth simply because without categories it is not shown there. Women’s clock – the increase has become. Sports clock – growth. Well, we have not very accentuated children’s watches. Fitness braids. With fitness bracelets was the most difficult situation. If you look here, neither smartshopping nor performance was very difficult to cope with them. Well, unisexes need to be watched. Sports have grown. What’s here? Children’s clock also grew up.
Well, that is, if here, there was definitely growth in this category here. And this growth was very actively observed in smartshopings. Now, if you look, it was in smartwitching that it was very actively observed in some of the majority cases. There, if you look at different periods. Well, for example, you can take the same period of this August. Or not even August. Here is July. At first we had tracks in smartshoping.
And yes, well, by the way, it will be seen, this is the period when we very much increased traffic. So far I will show you. There was a period when we very much increased traffic from comrades. God, how to say. It’s July. Here is August. Yes. What period did I choose? Now I will show. July. November 4. We need not on November 4. Let’s choose July. Compared to the previous period. Here you can see how smartshoping asked in some conversions. Here, 42%, 55%are visible here.
And guess why? Because the same Performance Max was just activated here. That is, here if you look, here July for a month we really began to grow conversions. During this period, smartshoping sagged. Therefore, we reduced our appetites here. And then it became a little different. Now let’s see in August the month. Well, there is already some, for example, there is a male fall, but it is not as much or actively, as it was, for example, in previous periods.
Well, for example, the category of male times here is 10% of all. Although here without category, for example, an increase of 20%.
For example, here the smart mines have sank with us, because the performance reduced its activity and, accordingly, smartshoping reduced its activity. That is, there was a direct very clear correlation. If we increased very actively on performances, then a slightly lifted smartshoping. As soon as we in some category were not very actively promoted in performances, I immediately sagged smartshoping. And such a correlation can be found a lot, such nuances, factors, which, here conditionally, prevented this whole story from being somewhat more or less stable.
Since our task was to make it more or less adequate, stable and adequate. Accordingly, it was in competition. We were, conditionally, an additional additional income to smartshoping, conditionally. And here, after this period, when the smartshoping was conditionally removed, it is clear that traffic was completely migrated here with conversions directly to Performance Max.
Here, for example, these two campaigns are even visible. Here on the base of the lower smartshoping was created. And this one has been created for a long time. And if we take this period, for example, on October 11, there is a jump here. Now, if you look, this one gives less conversion.
profitability is so a little better. Here, the price for the conversion is almost identical. This profitability gives, because here the check is slightly higher. The price of a click is almost identical. But this one already spends much less.
Well, the client will be planted here. All this should be followed, what is happening there. But still it is very noticeable. And if you take, for example, conditionally, then 3000 total sales. Half of them makes one performance, everything else, for example, does this structure, which the client has yet been modified there, for example, for himself personally. Something works well for this, something works worse for this. But, conditionally, you should always work conditionally with this. It will not convert so simply in itself.
What was tested in the project and what is the focus on?
Yes, let’s now move on to what has been tested here. And what would I, in principle, continue to continue to focus on the site of the client. But of course, if the client wants to further plant these advertising accounts. To whom this is not a secret, not all agencies require blackmail, forced to stay in a booty. That is, you can go in, optimize, then accompany all this yourself.
If we take a situation when, so that I still reproach, this is necessarily a breakdown in brands. Because she partially worked out with the client and had to extend it. Because in fact there are many brands that googled straight. Second, it is imperative to continue breaking through brands or categories, because it has men’s standard ones there, there are military and other hours.
The third is necessarily testing at prices. Continue a breakdown of a clock at the price of the clock, because it has very cool top models of a more budget price range and there are cool models of a more expensive check. Those who sell some very premium products, there is a watch and so on, your task is possible if there is something budget or to spread and test this whole story in brands, because it really works in your case.
On the target price for the conversion
The target price for the conversion, by the way, very cool knocked down all the nuances that were on the build-up on profitability. That is, since we worked with duplicates of FIDS, very often these performances broke a little in terms of a knock on the shows. And somehow the target price for the conversion allowed a little to bring down the previous training and continue to transfer again to all profitability.
At the target price for the conversion, with a given price range, it practically did not work out. It was directly units of sales, regardless of the budget, it could be two, three, two or three sales.
Donasting proposals for buying tag
The next thing that would have done on the site of the client here, we had a bustro -building tag in this case over time. But still there was a certain percentage of the return of these goods. It was very relevant for the client to track the real actual profitability within the framework of the campaign. Therefore, pay attention to the first that you will not immediately load adequate profitability either in analytics or Google Ads.
That is, you, to understand, within 72 hours, if you see a super poor result there, exclude advertising campaigns, retroactively, such a situation may arise that everything was pretty good there. This has been held more than once on this advertising account, when we excluded some companies, transformed. And they retroactively then after a certain period saw that they were very good profitability.
The following, if you have a lot of returns or you really work within your payback, profitability that is beneficial to you, then your task will send the correct conversions of the order by the API Google ADS. This is a separate story called offline conversion.
If you have your own developer, your own programmer who can write this whole story under you will be the best option. Because most CRM systems, they still do not have direct integration with the fourth analytics, not to mention the Google ADS API. Because a lot of different CRM systems of the systems cannot adapt to all possible platforms. But if you have an intelligent programmer, then please do it.
Use to track email conversions, not phones
Another cool thing, which is still on the site of the client, although he refused, protested. We had it on the call with carbon. The fact that for expanded tracking of conversions use e-miles, not phones, this is mandatory. E-miles are a big user behavior analytics on the site, for example, your phones will do. Google does not know how to work with phones.
Therefore, when you add the database of phones, it is absolutely non -working. But the base of e-male gives at least something. When you turn on the expanded tracking of conversions, the E-male base is very cool. Even from the very beginning, she begins to work out pretty well.
These are all sites, all platforms are also held here. And having a class programmer that is now accompanied by this project, I would still configure the sending of the right conversions to Google ADS so that advertising campaigns can already work on profitability, which is already stretching from real sales. Because some of the campaigns in different periods had different variability, the totality.
No need to focus on the signals
Another cool thing is that it is not necessary to concentrate attention very on the signals. This again, this project confirms that the presence of signals can even limit you. Because here was one of the first campaigns, which was launched on remarketing lists. And she converted very little. And only on wide targeting did she begin to convert it better.
Pay attention whether the ID of the goods were correctly indicated
The following, if you play with duplicates of feeds, then do not think that if you have done it for different campaigns, then it will work positively. No, on the contrary, it is even clear here that as soon as the campaigns, they received their native ID shniki. Here is a performance, look at what immediately in growth.
Just the traffic began to migrate back now completely in Performank. There is no competition and so on, because it works quite well. And this Performance is much more lively than, for example, it can exist longer than these performances working on non-native IDshniks.
Recommendations based on it shown on the case
God, Fuh, detained you for a very long time with some kind of conversation. This is one of these elements, so interesting cases when you need to keep the balance in order not to spoil what was here. Because there is still a search. There is still a disc. Here you can still shoot the video on the example of this case. Here, still a performance can be decomposed into a certain part of the cases.
The essence is? The bottom line is this case? The point is that as long as you have an element that works well enough, use it while you can. Second, if you use any automatic strategies, especially Performance Max, give it all the same data. Even on Shopify, when different IDs are recorded in the feed and on the website. This is also not ok and you need to stabilize them in order for a very beautiful picture to emerge; this has already been done on more than one project.
Next. Try to grow your performances into very significant budgets so that they are large enough. For example, this one is 18,000 UAH. Here, for example, it is based on smart shopping. These are the same settings so that old performances can be made in smart shopping.
Next, try, test different structures. Will it be brands, or will it be price breakdowns, or will it be strategy breakdowns. Constantly test something new in your advertising account. Try different groupings, try to intersect different product categories. See who buys your products, men, women, what specific age categories or any other parameters. In order to understand how you can further segment your products.
The best option here is to always draw the client’s attention to the fact that you have displayed products that are not shown. The best option is to try to display products that are not simply shown in impressions.
Well, the only nuance that broad queries sell here. Therefore, you need to spend a little time on this to bring them to the shows. Because they compete a little for requests when there are, for example, twenty great products that convert, or even a hundred products convert well. There are also those at the end of the list who do not show up at all. Here you need to forcefully prioritize them in search results.
Conclusion
So, in principle, this is such a cool beautiful case. I think I will report something else from these cases in order to show how it works. And so that I have some kind of aggregation of some useful tricks or life hacks for you if you plan to use them in your work.
Well, plus, this is a demonstration of real things that I talk about on YouTube and how they are translated, for example, into reality. What interferes, what does not interfere. Some real problems or troubles in accounts that any of us can constantly experience. Because we are all human and especially Google, which also depends on people and the work they do on these advertising tools. Therefore, in principle, I will finish it. And I hope that I will soon post a couple more cases on the same project.
