Hi everyone, my name is Jana and I am a Google logistician. I am in charge of delivering targeted business audience with the right parameters. I scale online sales of businesses globally using Google Ads tools.
Description of the case study that will be discussed – what did the client come in with?
Today’s video will also include a case study of the team. Our topic is dietary supplements. Market – Ukraine. Targeted CPA-ka we go, the price per conversion up to 250 hryvnias.
The client has only had a search before. You can even look, you can see how long or short the search was there. And the price per conversion is definitely from search.
How many calls and sales will I get by ordering contextual advertising from you?
You can see that search was making a pretty decent price. You can see the number of conversions here. We added more additionally on Performance Max, so if you run only search, don’t forget that merchandise, in principle, can give you quite a positive result.
Difficulties encountered
What were the challenges we faced? The fact that the price per conversion was not always good enough. We have been gradually collecting conversions. I would not say that this is a super cool result, but we have some relative plus or minus scaling, traffic penetration and stabilization in terms of price per conversion.
Because it’s better to collect cheaper price per conversion in BADs in the statistical significance of analytics to make it easier to scale further. Why? Because the bulk of conversions come from Pmaxes. Pmaxes in this case are such a rather capricious type of advertising campaign. A lot of Googlers are making changes to them now, and they learn faster. You can see that immediately by the algorithms they use. And you can also see that they get to the values you need faster. And the values you want are usually what? It’s of course either the price per conversion or the profitability that we want to use.
The work the team has done in the first two weeks
Why else would I want to show you this case? To show that not everything is so smooth. You see this period where we didn’t fit into the conversion rate at all while we were teaching the AI thing to do a certain result. We usually have the first two weeks of optimization weeks. What are these two weeks of optimization weeks, I call them? Because in this period we test structures, we test what is better in conversion, what is worse. Which products can scale, fit into KPIs, which ones cannot scale. Or, for example, Conversion rate needs to be adjusted for them so that they can get those conversions. Maybe edits to feeds, testing structures, strategies, and so on. In principle, during these two weeks, we usually find a scaling point, a fulcrum in analytics, which we try to fix and scale even more in the next two weeks. Of course, it will be great if during this period we manage to stabilize the price per conversion and get it plus or minus within fairly narrow limits with a reserve. Because it doesn’t take much intelligence to eat all the margin or all the CPA-ka, but gradually scaling up and teaching the AI-thing to generate a cheaper price per conversion is usually a key factor, a key nuance when scaling any project.
The specifics of setting up advertising campaigns for dietary supplements
Dietary supplements, I remind you, are not a super easy niche for sales. There’s quite a bit of specificity high in the type of products that you’re selling. What this dietary supplement is used for, how competitive it is, what competitors are present in the auction, how aggressive they are, and so on. So, you have to keep a balance between different some factors here. So far, the growth in conversions has been so systematic.
How many calls and sales will I get by ordering contextual advertising from you?
If I remove all the graphs now, you can see how much the dynamics is going positive. Here is the stabilization going up in the number of conversions. I wouldn’t say that this is the maximum number of conversions you can get here. You could scale in units even faster. You can see how they are in steps. We can describe it as some kind of staggered scaling when we try to shoot through traffic and see how correct or not it will be.
Because Conversion rate also means a lot here. The higher it is, the more convenient it is to scale. It is not super high enough for us during this whole period. It is average, conditionally speaking, for the hospital. But it would be desirable that it would lead to some more stable high sizes. But for now, we’re in the first two weeks of stabilization and optimization. Then we will see if we can increase it.
The price per conversion, as you can see, in the first two weeks it was not shown by some fancy, beautiful analytics. Accordingly, it had to be handled carefully in traffic. Why? Because it’s very easy to teach Pmax to generate a higher price per conversion. Even though there was a slightly cheaper price per conversion here. Yes? 10 conversions at 61 hryvnias. But in general, there were situations when we had a rather high conversion rate. In order to stabilize it somehow, we had to work on the Conversion rate to stabilize it and make it more predictable. Accordingly, now this time period is more stable, as you can see. That is, he is also at a price per conversion of 149, for example, on March 18, which is such a smoother statistics analytics. Even in those intervals it’s 227, even though there aren’t that many sales.
Features of AI training in Performance Max
There’s a balance to be struck here. A balance of what? You have to realize that AIs are learning very fast right now. But learning how? Maybe it used to have some algorithm sewn into it more than some super… Let’s put it this way, more freedom was given to it in terms of scaling attempts. Now, most likely, the AI-thing has been squeezed within the framework of some sort of: learning mechanics, prediction model, predicted rate, conversion rate, where to show it, and so on. So, accordingly, it is now able to come more quickly to your values that you plan to put to it. That is, if before you could start with small profit margins and gradually scale them up, now you can set some profit margins. It will come to them very quickly in actual values. So it’s very important to give some or more median analytics that you can further deal with. That is, if it’s a disruption in some kind of price per conversion on your test versions of the campaign. Or already immediately teach on some smoother, slower dynamics of how they will get conversions.
Conclusions
In general, I’m showing this with the demonstration of conversion rates, effectiveness in order to once again confirm that the result that can be obtained here is not the result of some accidents, but it is the result of certain regular actions in the advertising account. Therefore, accordingly, it affects the overall picture that we have here now. That is, if you stabilize, if you make sure of the size of this result, if you continue to stabilize it, then further this mechanics of movement, such steps conditionally, further continues only for a greater number of conversions.
That’s a little case study. I think we will continue to work with the client. Next we will try to make even more conversions, which could be stable not only in volume, because we have brought it up to five pieces of conversions steadily.
But the price per conversion was not stable everywhere. We have stabilized the price per conversion, now the task is to scale the number of pieces of conversions. Here it’s eight, now it’s ten. So there’s no drawdown to a five like there used to be. This is the case I wanted to show you today.
