Press "Enter" to skip to content

Which Conversion Attribution Model Should I Use?

When you’re setting up a conversion action, you’re faced with the choice of which attribution model to use. The conversion action measures a particular action that somebody took on your website after clicking on your ad. The types of conversions may be a call, a purchase, a filled-out opt-in form, and other things. 

The attribution model will tell Google how you want to keep track of these conversions.

If somebody clicked on more than one of your ads, you need to tell Google which of those clicks should get credit for the conversion. There are different weighted options where certain clicks get more credit for each conversion. But essentially, the attribution tells Google how to give credit for those clicks.

First, let me say that Google will take credit for every click. If you originally generated a customer or click from Facebook, and then, later on, they click on one of your Google Ads, Google doesn’t care about Facebook or any other time that that customer has been to your website. Google will take credit for that click, no matter what

This only applies to the clicks that come from Google Ads. 

There’s one attribution model, that I normally use. But I’m going to talk about all the attribution models. I’m not saying that you have to use the one I prefer, and that’s why you should be familiar with how each of these models work. Whichever model you’re using, you should know how it’s going to impact your campaign.

By the way, all of these only matter if there are multiple clicks. Most of the times, that’s not the case. Usually, there’s only one click, and then either the person becomes a customer or not. You’ll actually find that by using different models, you won’t find much variation in your actual results because most of the time it’s just one click that leads to the sale.

Attribution Models:

1)  FIRST CLICK

It gives all the credit for the conversion to the very first click. 

The first time the user comes to your website, maybe they don’t buy or become a lead right away. Later on, they click another one of your ads and end up buying. In this case, only the first click is going to get credit for that sale. 

The reason why I always use this attribution model is this:

When I’m optimizing campaigns, I’m most concerned with that first click, with the initial campaign that’s targeting the cold traffic that’s coming to a website. I’ll also be running retargeting campaigns because I know for sure that they are worthwhile. I don’t really need to see conversions in those campaigns to know that they’re working.

What’s more important for me as a campaign manager is to know where that cold traffic is coming from and how well that cold traffic is converting. In this way, I can optimize the campaign to get more cold traffic. Because that’s what moves the needle forward more than anything, it’s being able to get more cold traffic into a website which eventually leads more people to becoming a customer.

By selecting FIRST CLICK, I’m optimizing only for cold traffic. With that in mind, I need to understand that if I’m running remarketing campaigns, I’m not really going to be seeing conversions in those campaigns. There will be some, because if people came from other traffic sources and then I’m remarketing to them on Google Ads, (perhaps the first click in Google Ads is in a remarketing campaign).

But if their first click on Google ads is to a cold traffic campaign, that’s where the conversion is going to show up. It won’t be in my remarketing campaign, so the results will be skewed. I’m fine with that. It’s more important to be able to optimize for that first click, that cold traffic coming in. This is why I use FIRST CLICK.

2) LAST CLICK

In this case, all of the conversion will be credited to the last click that happened before the customer became a conversion. There won’t be any conversion showing up in the first click. 

3) LINEAR

It’s going to evenly distribute the conversion across multiple clicks. This is where we start to see decimal places for conversions. If there’s one conversion and there were two clicks, each of those is going to count as .5 of a conversion. If there were four clicks that happened before a conversion, each of them is going to count as .25 of a conversion with this linear model.

4) TIME DECAY

In this model, the clicks that happened closest to the time of conversion are going to get more weight. There is a formula that they use for this. But all you need to know is that the very first clicks get very minimal weight. The last couple of clicks before the conversion happened, are the ones to get weighted more heavily. You might see a .2 conversion on 1 click, and a .8 on another click.

5) POSITION BASED 

This gives 40% credit to both the first and last click, and then the remaining 20% is distributed across all the other clicks in the path. If you had 3 clicks, the first click is .4, the 2nd click gets .2, the last click also gets .4. 

6) DATA DRIVEN 

This is an option that’s not available to everybody. There has to be a good amount of data in your account for Google to open this up to you. This uses Google’s artificial intelligence to give credit to the click that Google feels deserves most of the credit. Here, you are relying on Google’s system to know which click contributed more to the conversion.

In my opinion, this is not too helpful. I would use any of the previous options rather than this because there is no real way to know how accurate this is.

What is the Bottom line?

The attribution model that I am almost always using is FIRST CLICK, and that’s so I can optimize more heavily for cold traffic. You can use any of the attribution models mentioned. The most important thing is making sure you know how everything works!