Conversion Paths & Different Attribution Models

What are conversion Paths Why they Matter?

When we run ads on digital, we use multiple platforms like – Google, Facebook, Email, SEO, etc. Hence giving customers the convenience to interact with the brand anywhere.

But then, how do we understand which platform influenced what in the customer buying journey? And more importantly for D2C brands which platforms generated awareness and which got conversions.

For the ease of understanding let’s take conversion to be our goal. Now we run conversion ads on Facebook, Google, customers can come to the website and purchase organically or via email. 

So then how can we understand which channels the customer interacted with throughout their journey until they made purchase, how many times did they interact and finally made purchase from which platform.

Google Analytics gives us a picture of the top conversion paths. Here is a snapshot of the same: 

 

 

We can see this report in Google analytics under Conversion -> Multi-channel Funnels - > Top Conversion Paths. Why don’t you open this right away and check it for your brand?

This will help us understand which paths are working for us and how the customer is interacting with your brand website.

It will help us answer important questions like –

1. What is the customer buying journey?
2. How many times the customer is interacting before making the purchase.
3. Which platform is performing good for which part of the customer buying journey (like Awareness, Conversion, etc.)?
4. Which platform is helping us increase direct or organic contribution?

Once we have this information, now we want to understand which platform or touchpoint is getting the attribution of conversion. This is where Attribution models come into picture.

 

What are Attribution Models and Why they Matter?

Attribution models gives us control to assign credit to different marketing touchpoints, that the customer interacted with. This helps us analyse the performance of each channel properly and take informed calls on how we want to use different platforms to increase our sales.

There are different attribution models that we can use. Google Analytics also has attribution reports which can help us analyze this for our website.

To understand different types of attribution models, let us take an example of conversion path:

Google Search -> Facebook -> Direct

Now, here the customer came to the website from Google Search first, then through Facebook (probably a retargeting campaign) and then finally made the purchase by coming to the website Directly.

Different Attribution Models: 

1. First Click: In this model, sale is attributed to the first touchpoint. We can use this model to see which platform is getting us new customers or visit to the website. In the example above, if we use first click model 100% attribution of the purchase will be given to “Google Search”.

2. Last Click: If we use this model then, sale is attributed to the last touchpoint. We can use this model to see which platforms is helping us get conversions. This is also the default model in most analytics tools and is used widely. In the example above, if we use Last click model 100% attribution will be given to “Direct”. 

3. Linear: Here, the sale is attributed to all the touchpoint equally. In the above example, 33% attribution will be given to each of the 3 touchpoints.

4. Data Based: There are various Data based models like Time Decay, U-shaped, V-Shaped, etc. The sale is attributed to all the touchpoints but not equally. But we don’t need to get into those details as the most used model that is also good for small D2C brands is the Last Click or Linear Model.

Now if we are using Last Click model by default. Then, while analysing the platform performance we should definitely look at other attribution models like first click and Linear to understand if any platform is assisting another platform, which is performing well as per the last click. This is called “Assisted Conversion”.

Let’s take another example:

Google Search -> Facebook

Now here if we have default Last Click Attribution Model then the attribution of all the sales will happen to “Facebook”. In this case a marketer can take the call to reduce budget or altogether stop google search campaigns believing that they are not performing. And increase the budget on Facebook.

But will this decision be correct?

It will actually reduce the performance of Facebook and mess up the ROAS.

Analysing attribution models helps us answer important questions like:

1. Which platforms are helping us get awareness and new visitors which are converting later? (Use First Click model for this)
2. Which platforms are helping in the final conversion? (Use Last Click model for this)
3. How to better allocate Marketing Budget?
4. Which platforms are working in together as top conversion paths?

Now let’s look at a sample report from google analytics. We can find the same under Conversions -> Multi-channel Funnel -> Assisted Conversions.

Here, we can clearly see that Organic is contribution 244 conversions but is assisting in 1306 conversions, hence we can confidently say that SEO is working for us. It is helping people discover our brand website and getting us new traffic and customers.

Although people are not converting immediately, but we can retarget them and eventually try to get conversions.

There is another very important report in Google Analytics, which helps us compare the data for 2 different conversion models. We can find the same under Conversions - > Multi-channel Funnels -> Model Comparison Tool.

Let’s look at a sample screenshot. Here I have selected Last Click and First Click. We can select up to 3 models and compare.

 

 

We can see all these reports at campaign level also and do the attribution analysis at source/medium or campaign level. I would highly recommend that we do this.

Attribution analysis should be done once a month or once every two months. It will help us plan the next month better and allocate budget wisely. It will also help us understand the situation of our business. This analysis not only helps us answer many important questions but also give us food for thought.

We can run experiments to see if we can change the customer behaviour in our favour. Or we can check how adding new platforms/ campaigns in our marketing strategy is helping in the customer journey.

Note – All the screenshots and data here is dummy and is not accurate.