When it comes to attribution, we often fixate on that last click. But, does this perspective truly paint the full picture?
Ignoring it may seem tempting, but it won’t guide you in making informed decisions about budget allocation and investment strategies.
Attribution is, in essence, a unique lens through which we view our marketing efforts. The key lies in selecting the perspective that aligns best with your customers’ behavior and business model. Unlike Universal Analytics, Analytics 4 empowers us not just to pick an attribution model but to also make meaningful data comparisons when employing different models. In this video, we’ll demystify the process – from model selection to understanding their implications and conducting insightful comparisons.
The video was recorded in Hebrew, but you’ll find English subtitles:
Ready to take the next step in optimizing your attribution model and making data-driven decisions?
Schedule a free introductory call with us today to explore how we can help your business thrive.
Here’s a transcription of the video:
Is attribution nonsense? What is attribution anyway?
Attribution is a credit we give for a conversion that happened on the site.
Many people think that attribution is the last click, but it is a mistake to think that way.
Attribution is actually a point of view on your marketing activity.
And the trick is to choose a point of view that suits your business, the behavior of your customers, and your business model. This will allow you to make smart decisions about where you should invest the money in marketing and advertising.
Analytics 4 allows you to choose the appropriate model.
In this video, I will show you how to do it.
Shalom, my name is Nathan Kreiderman, I am the founder of analyzit.
What we do is take complex data and turn it into conclusions and profits for the business.
In Analytics 4, we can define the attribution model we want to work with.
And we will do it from the admin. Let’s go to gear here.
And under properties, we will find attribution settings. We will click on it and we can choose one of the models that Analytics offers. Let’s see what these models are.
In the end, we will talk about data-driven. Let’s start from the Last Click that you all know. It means that the conversion will be related to the last channel from which the user came. First Click is the opposite – the conversion will be associated with the first channel from which the user came from.
Linear – this means that the value of the conversion will be divided equally between all the channels that the user touched.
Then there is Position Based. This means the conversion will split among the channels the user came from when the first and the last touch will get more, and all the rest will be divided between them.
Time Decay – this is as closer to the date of the conversion we get, to the date of the order, it will receive more value, more attribution. This is called Time Decay. For example, it is very suitable for businesses that deal with hotels or travel, flights. As I get closer to the time when I need the hotel, the value of this channel will increase.
A new one that Analytics 4 recently added, and this is what is called last click ads preferred, which is very similar to what we had in Analytics Universal. It was called last non-direct. This means that the conversion will be associated with the last channel it came from, which is neither a direct channel nor an organic channel. That means a channel that has UTM. That’s how we looked at it until now at Universal.
There is one more new model here called data-driven, and it appears as Google’s recommended option. As far as I’m concerned, it’s not that recommended, but everyone makes their decision, and basically Google says – we have some kind of AI here. We will activate some kind of robot, some kind of algorithm that will take all the entries of the user, will consider and associate the conversion in some way by an algorithm. That’s the advantage here. Basically, Google tells us – let us invest for you, and we will make some kind of model that is very sophisticated together with AI. And the downside of it leaves us pretty much in the dark; we don’t really know what this model is and how Google does it, and we can’t be sure that it fits our business.
In the end, the most important thing here is to choose the model that suits your business best. Even if you have decided on a certain model, it is always worthwhile to look at another model through another point of view. In the end, as we said – attribution is just a point of view.
So how I will do it in Analytics 4? I will access the advertising reports.
And under advertising reports, I have a report called Model Comparison, which actually allows me to compare different models.
So what do I see here?
In this report, I have two attribution models that I can choose from. For example, here it is Last Click against Data-Driven. I’m actually comparing between these two models. I can see what would happen, which channel would get more or less conversions, and also the value of conversions, and especially revenues. If I were looking at this channel according to the different models.
So, let’s take, for example, paid search, and you can see if I looked at Last Click, I would get $426, and if I looked according to Data-Driven, I would receive only $50, which is 80% less in Data-Driven compared to Last Click.
Let’s compare Last Ads Preferred; I wonder how it will look. There is not much difference. Let’s try Last Click. Not much difference in this respect, only 7%. But let’s look at First Click. Yes, if I look at the First Click, you can see that Data-Driven shows minus 88 percent here, and you can see that most users are not exposed to the site from the search but come to the search after they have already been exposed from another source. It may be that they come here to search with the help of a brand expression, and therefore in a driven data or in a model of First Click, it shows us the revenue that will be associated with the channel; this will be much, much, much lower than Last Click.
In this way, I can compare models and again, choose the right model for the business. And also to look