Understanding Conversion Credit distribution for Attribution Models in Google Analytics

Following article is an excerpt from my best selling book: Attribution Modelling in Google Analytics and Beyond which I am sharing it here for the first time, to benefit the wider audience:

In the present article I will show you, how Google Analytics distribute conversion credit to various touch points in a conversion path for different attribution models.

This insight will greatly improve your understanding of various GA attribution models.

But before we move forward, there is one caveat.

This article is bit advanced.

It is advanced in the sense that, it assumes that you already have a working knowledge of attribution modelling in GA.

Attribution Modelling itself is quite an advanced topic in GA for many but the present article takes this subject to a new level.

I won’t be explaining the very basics of attribution modelling in the present article.

So if you are brand new to attribution modelling then read this article first: Beginners Guide to Google Analytics Attribution Modeling.

In order to get the most out of the present article make sure you understand:

Conversion Credit Calculations for Last interaction attribution model

The ‘Last interaction attribution model’ (also known as last touch attribution model) assign 100% credit for conversion to the last interaction in a conversion path.

Google Analytics uses this model by default for Multi channel funnel reports.

Following example will help you in understanding how the conversion credit is calculated in case of last interaction attribution model.

Consider following conversion path of ‘path length’ 4:

conversion-path1

This conversion path can also be represented by the following table:

table1

Here the user was exposed to various campaigns in the following order before the conversion occurred on your website:  Campaign ‘A’ > Campaign ‘B’ > Campaign ‘B’ > Campaign ‘C’

Under last interaction model, the last touch gets 100% credit for conversion. So conversion credit distribution would look like the one below:

distribution1

Here,

Campaign ‘A’ gets 0% credit for conversion.

Campaign ‘B’ gets 0% credit for conversion.

Campaign ‘C’ gets 100% credit for conversion.

Get the Free E-Book (52 Pages)

Consider second conversion path of path length 3:

conversion-path2

This conversion path can also be represented by the following table:

table2

Under last interaction model, the last touch gets 100% credit for conversion.

So conversion credit distribution would look like the one below:

distribution2

Here,

Campaign ‘A’ gets 100% credit for conversion.

Campaign ‘B’ gets 0% credit for conversion.

Campaign ‘C’ gets 0% credit for conversion (as campaign ‘C’ is not in the conversion path)

So total conversion credit for campaign ‘A’ so far = (Conversion credit for campaign ‘A’ in the first conversion path + Conversion credit for campaign ‘A’ in the second conversion path) /2 = (0% + 100%) / 2 = 50%

Total conversion credit for campaign ‘B’ so far = (Conversion credit for campaign ‘B’ in the first conversion path + Conversion credit for campaign ‘B’ in the second conversion path) /2 = (0% + 0%) / 2 = 0%

Total conversion credit for campaign ‘C’ so far = (Conversion credit for campaign ‘C’ in the first conversion path + Conversion credit for campaign ‘C’ in the second conversion path) /2 = (100% + 0%) / 2 = 50%

Consider third conversion path of length 2:

conversion-path3

This conversion path can also be represented by the following table:

table3

Now the conversion credit distribution would look like the one below:

distribution3

Here,

Campaign ‘A’ gets 0% credit for conversion.

Campaign ‘B’ gets 100% credit for conversion.

Campaign ‘C’ gets 0% credit for conversion.

So total conversion credit for campaign ‘A’ so far = (Conversion credit for campaign ‘A’ in the first conversion path + Conversion credit for campaign ‘A’ in the second conversion path + Conversion credit for campaign ‘A’ in the third conversion path) /3 = (0% + 100% + 0%) / 3 = 33.33%

The total conversion credit for campaign ‘B’ so far = (Conversion credit for campaign ‘B’ in the first conversion path + Conversion credit for campaign ‘B’ in the second conversion path + Conversion credit for campaign ‘B’ in the third conversion path) /3 = (0% + 0%+ 100%) / 3 = 33.33%

The total conversion credit for campaign ‘C’ so far = (Conversion credit for campaign ‘C’ in the first conversion path + Conversion credit for campaign ‘C’ in the second conversion path + Conversion credit for campaign ‘C’ in the third conversion path) /3 = (100% + 0% + 0%) / 3 = 33.33%

If we assume that there are only three conversion paths in our selected time period, then the sum of conversion credit for campaign ‘A’, ‘B’ and ‘C’ under last interaction attribution model would be =  33.33% + 33.33% + 33.33% = 99.99% (rounded off to 100%)

Since the sum of the conversion credit for three campaigns is 100%, this prove that my conversion credit distribution calculations are all correct.

Conversion Credit Calculations for First interaction attribution model

The ‘First interaction attribution model’ (also known as First touch attribution model) assign 100% credit for conversion to the first interaction in a conversion path.

Following example will help you in understanding how the conversion credit is calculated in case of first interaction attribution model.

Consider following conversion path of ‘path length’ 4:

conversion-path1

Under first interaction model, the first touch gets 100% credit for conversion. So conversion credit distribution would look like the one below:

distribution1

Here,

Campaign ‘A’ gets 100% credit for conversion.

Campaign ‘B’ gets 0% credit for conversion.

Campaign ‘C’ gets 0% credit for conversion.

Consider second conversion path of path length 3:

conversion-path2

Under first interaction model, the first touch gets 100% credit for conversion. So conversion credit distribution would look like the one below:

distribution2

Here,

Campaign ‘A’ gets 0% credit for conversion.

Campaign ‘B’ gets 100% credit for conversion.

Campaign ‘C’ gets 0% credit for conversion.

So total conversion credit for campaign ‘A’ so far = (Conversion credit for campaign ‘A’ in the first conversion path + Conversion credit for campaign ‘A’ in the second conversion path) /2 = (100% + 0%) / 2 = 50%

Total conversion credit for campaign ‘B’ so far = (Conversion credit for campaign ‘B’ in the first conversion path + Conversion credit for campaign ‘B’ in the second conversion path) /2 = (0% + 100%) / 2 = 50%

Total conversion credit for campaign ‘C’ so far = (Conversion credit for campaign ‘C’ in the first conversion path + Conversion credit for campaign ‘C’ in the second conversion path) /2 = (0% + 0%) / 2 = 0%

Consider third conversion path of length 2:

conversion-path3

Now the conversion credit distribution would look like the one below:

distribution3

Here,

Campaign ‘A’ gets 100% credit for conversion.

Campaign ‘B’ gets 0% credit for conversion.

Campaign ‘C’ gets 0% credit for conversion.

So total conversion credit for campaign ‘A’ so far = (Conversion credit for campaign ‘A’ in the first conversion path + Conversion credit for campaign ‘A’ in the second conversion path + Conversion credit for campaign ‘A’ in the third conversion path) /3 = 100% + 0% + 100%) / 3 = 66.67%

The total conversion credit for campaign ‘B’ so far = (Conversion credit for campaign ‘B’ in the first conversion path + Conversion credit for campaign ‘B’ in the second conversion path + Conversion credit for campaign ‘B’ in the third conversion path) /3 = (0% + 100% + 0%) / 3 = 33.33%

The total conversion credit for campaign ‘C’ so far = (Conversion credit for campaign ‘C’ in the first conversion path + Conversion credit for campaign ‘C’ in the second conversion path + Conversion credit for campaign ‘C’ in the third conversion path) /3 = (0% + 0% + 0%) / 3 = 0%

If we assume that there are only three conversion paths in our selected time period, then the sum of conversion credit for campaign ‘A’, ‘B’ and ‘C’ under first interaction attribution model would be =  66.67% + 33.33% + 0% =100%

Since the sum of the conversion credit for three campaigns is 100%, this prove that my conversion credit distribution calculations are all correct.

Conversion Credit Calculations for Linear attribution model

The ‘Linear attribution model’ assign equal credit for conversion to each interaction in a conversion path.

Following example will help you in understanding how the conversion credit is calculated in case of linear attribution model.

Consider following conversion path of path length 4:

conversion-path1

Under linear attribution model, the conversion credit for each interaction would be calculated as: 1/ conversion path length.

Since the length of our conversion path is 4, so conversion credit for each interaction would be calculated as: 1/4 = 25%

distribution1

Here,

Campaign ‘A’ gets 25% credit for conversion.

Campaign ‘B’ gets 50% credit for conversion.

Campaign ‘C’ gets 25% credit for conversion.

Consider second conversion path of path length 3:

conversion-path2

Under linear attribution model, the conversion credit for each interaction would be calculated as: 1/ conversion path length.

Since the length of our conversion path is 3, so conversion credit for each interaction would be calculated as: 1/3 = 33.33%

distribution2

Here,

Campaign ‘A’ gets 66.66% credit for conversion.

Campaign ‘B’ gets 33.33% credit for conversion.

Campaign ‘C’ gets 0% credit for conversion.

So total conversion credit for campaign ‘A’ so far = (Conversion credit for campaign ‘A’ in the first conversion path + Conversion credit for campaign ‘A’ in the second conversion path) /2 = (25% + 66.66%) / 2 = 45.83%

Total conversion credit for campaign ‘B’ so far = (Conversion credit for campaign ‘B’ in the first conversion path + Conversion credit for campaign ‘B’ in the second conversion path) /2 = (50% + 33.33%) / 2 = 41.67%

Total conversion credit for campaign ‘C’ so far = (Conversion credit for campaign ‘C’ in the first conversion path + Conversion credit for campaign ‘C’ in the second conversion path) /2 = (25% + 0%) / 2 = 12.5%

Consider third conversion path of length 2:

conversion-path3

Since the length of our conversion path is 2, so conversion credit for each interaction would be calculated as: 1/2 = 50%

distribution3

Here,

Campaign ‘A’ gets 50% credit for conversion.

Campaign ‘B’ gets 50% credit for conversion.

Campaign ‘C’ gets 0% credit for conversion.

So total conversion credit for campaign ‘A’ so far = (Conversion credit for campaign ‘A’ in the first conversion path + Conversion credit for campaign ‘A’ in the second conversion path + Conversion credit for campaign ‘A’ in the third conversion path) /3 = (25% + 66.66% + 50%) / 3 = 47.22%

The total conversion credit for campaign ‘B’ so far = (Conversion credit for campaign ‘B’ in the first conversion path + Conversion credit for campaign ‘B’ in the second conversion path + Conversion credit for campaign ‘B’ in the third conversion path) /3 = (50% + 33.33% + 50%) / 3 = 44.44%

The total conversion credit for campaign ‘C’ so far = (Conversion credit for campaign ‘C’ in the first conversion path + Conversion credit for campaign ‘C’ in the second conversion path + Conversion credit for campaign ‘C’ in the third conversion path) /3 = (25% + 0% + 0%) / 3 = 8.33%

If we assume that there are only three conversion paths in our selected time period, then the sum of conversion credit for campaign ‘A’, ‘B’ and ‘C’ under linear attribution model would be =  47.22% + 44.44% + 8.33% = 99.99% (rounded off to 100%)

Since the sum of the conversion credit for three campaigns is 100%, this prove that my conversion credit calculations are all correct.

Conversion Credit Calculations for Position based attribution model

By default, the ‘Position based attribution model’ assigns 40% credit to the first interaction, 20% credit to the middle interactions and 40% credit to the last interaction.

Following example will help you in understanding how the conversion credit is calculated in case of position based attribution model.

Consider following conversion path of ‘path length’ 4:

conversion-path1

Under position based model, both the first interaction and last interaction gets 40% credit for conversion and the remaining 20% credit is equally divided among middel interactions.

So conversion credit distribution would look like the one below:

distribution1

Here,

Campaign ‘A’ gets 40% credit for conversion.

Campaign ‘B’ gets 20% credit for conversion.

Campaign ‘C’ gets 40% credit for conversion.

Consider second conversion path of path length 3:

conversion-path2

Under position based model, the conversion credit distribution would look like the one below:

distribution2

Here,

Campaign ‘A’ gets 60% credit for conversion.

Campaign ‘B’ gets 40% credit for conversion.

Campaign ‘C’ gets 0% credit for conversion (as campaign ‘C’ is not in the conversion path)

So total conversion credit for campaign ‘A’ so far = (Conversion credit for campaign ‘A’ in the first conversion path + Conversion credit for campaign ‘A’ in the second conversion path) /2 = (40% + 60%) / 2 = 50%

Total conversion credit for campaign ‘B’ so far = (Conversion credit for campaign ‘B’ in the first conversion path + Conversion credit for campaign ‘B’ in the second conversion path) /2 = (20% + 40%) / 2 = 30%

Total conversion credit for campaign ‘C’ so far = (Conversion credit for campaign ‘C’ in the first conversion path + Conversion credit for campaign ‘C’ in the second conversion path) /2 = (40% + 0%) / 2 = 20%

Consider third conversion path of length 2:

conversion-path3

Under position based model, Campaign ‘A’ gets 40% credit for conversion and campaign ‘B’ gets 40% credit for conversion.

Since there are no middle interactions, the remaining credit of 20% is equally divided about campaign ‘A’ and campaign ‘B’.  

So,

  • Campaign ‘A’ gets 50% credit for conversion.
  • Campaign ‘B’ gets 50% credit for conversion.
  • Campaign ‘C’ gets 0% credit for conversion (as campaign ‘C’ is not in the conversion path)

The conversion credit distribution would look like the one below:

distribution3

So total conversion credit for campaign ‘A’ so far = (Conversion credit for campaign ‘A’ in the first conversion path + Conversion credit for campaign ‘A’ in the second conversion path + Conversion credit for campaign ‘A’ in the third conversion path) /3 = (40% + 60% + 50%) / 3 = 50%

The total conversion credit for campaign ‘B’ so far = (Conversion credit for campaign ‘B’ in the first conversion path + Conversion credit for campaign ‘B’ in the second conversion path + Conversion credit for campaign ‘B’ in the third conversion path) /3 = (20% + 40% + 50%) / 3 = 36.67%

The total conversion credit for campaign ‘C’ so far = (Conversion credit for campaign ‘C’ in the first conversion path + Conversion credit for campaign ‘C’ in the second conversion path + Conversion credit for campaign ‘C’ in the third conversion path) /3 = (40% + 0% + 0%) / 3 = 13.33%

If we assume that there are only three conversion paths in our selected time period, then the sum of conversion credit for campaign ‘A’, ‘B’ and ‘C’ under position based model would be =  50% + 36.67% +13.33% = 100%

Since the sum of the conversion credit for three campaigns is 100%, this prove that my conversion credit calculations are all correct.

Conversion Credit Calculations for Last non direct click model

The ‘Last non direct click model’ assign 100% credit for conversions to the last non-direct interaction in a conversion path.

Google Analytics uses this model by default for non-Multi channel funnel reports. 

Following example will help you in understanding how the conversion credit is calculated in case of last non direct click attribution model. Consider following conversion path of ‘path length’ 4:

conversion-path

This conversion path can also be represented by the following table:

table1

Under last non direct model, the last non-direct interaction gets 100% credit for conversion.

The last non direct interaction here is ‘Referral’.

So ‘Referral’ gets 100% credit for conversion.

The conversion credit distribution would look like the one below:

distribution1

Conversion Credit Calculations for Last Adwords Click model

The ‘Last Adwords Click model’ assign all the credit for conversions to the last Google Adwords click in a conversion path.

Following example will help you in understanding how the conversion credit is calculated in case of last Adwords click attribution model.

Consider following conversion path of ‘path length’ 5:

conversion-path

This conversion path can also be represented by the following table:

table

Under last Adwords click model, the last Adwords interaction gets 100% credit for conversion.

So the conversion credit distribution would look like the one below:

distribution

Conversion Credit Calculations for Time Decay attribution model

The ‘Time Decay attribution model’ assign more credit to the interactions which are closest in time to the conversion.

This model takes ‘half life in exponential decay’ into account while calculating conversion credit distribution.

Following is the formula to calculate half life in exponential decay:

half-life

Source: https://en.wikipedia.org/wiki/Half-life

Here,

N(t) denotes the quantity that has not decayed after a time ‘t’

N(o) denotes the initial amount before measuring decay. In our case this initial amount is 100% conversion credit.

‘t’ denotes the time span an interaction is away from conversion.

t ½ denotes half-life which is the time it takes for a quantity to reduce to half its original value.

By default, the time decay attribution model has a half-life value of 7 days. What that means, a interaction which occurred 7 days prior to conversion gets ½ the credit of the interaction that occurred on the day of conversion.  

Consider the following conversion path of path length 3:

conversion-path

Let us assume that interaction with position 1 occurred 14 days ago, interaction with position 2 occurred 7 days ago and interaction with position 3 occurred on the day of conversion.

Under time decay attribution model, the conversion credit for each interaction would be calculated as:

Campaign ‘A’ gets 85.7% (57.1% + 28.6%) credit for conversion.

Campaign ‘B’ gets 14.3% credit for conversion.

Campaign ‘C’ gets 0% credit for conversion.

The actual calculation of the conversion credit distribution in case of time decay model is quite complex and can’t be done or demonstrated manually. 

Conversion Credit Calculations for Data driven attribution model

Data driven attribution (DDA) model is an algorithmic attribution model.

It uses data modelling in which predictive algorithms are used to determine marketing touchpoints (both online and offline touchpoints) which should get more conversion credit than others.

Consequently, the conversion credit distribution for a DDA model cannot be demonstrated manually. 

Learn about the Google Analytics Usage Trends Tool

The Google Analytics usage trend is a new tool which is used to visualise trends in your Google Analytics data and to perform trend analysis.


Take your knowledge of Web Analytics to the next level. Checkout my web analytics training course.

Take your Analytics knowledge to the next level. Checkout my Best Selling Books on Amazon

Maths and Stats for Web Analytics and Conversion Optimization
This expert guide will teach you how to leverage the knowledge of maths and statistics in order to accurately interpret data and take actions, which can quickly improve the bottom-line of your online business.

Master the Essentials of Email Marketing Analytics
This book focuses solely on the ‘analytics’ that power your email marketing optimization program and will help you dramatically reduce your cost per acquisition and increase marketing ROI by tracking the performance of the various KPIs and metrics used for email marketing.

Attribution Modelling in Google Analytics and Beyond
Attribution modelling is the process of determining the most effective marketing channels for investment. This book has been written to help you implement attribution modelling. It will teach you how to leverage the knowledge of attribution modelling in order to allocate marketing budget and understand buying behaviour.

Himanshu Sharma

Certified web analyst and founder of OptimizeSmart.com

My name is Himanshu Sharma and I help businesses find and fix their Google Analytics and conversion issues. If you have any questions or comments please contact me.

  • Over eleven years' experience in SEO, PPC and web analytics
  • Google Analytics certified
  • Google AdWords certified
  • Nominated for Digital Analytics Association Award for Excellence
  • Bachelors degree in Internet Science
  • Founder of OptimizeSmart.com and EventEducation.com

I am also the author of three books:

error: Alert: Content is protected !!