This article is related to attribution modelling in Google Analytics. If you are brand new to attribution modelling, then read this article first: Beginners guide to Attribution Modelling in Google Analytics. 

What is the Google Analytics data driven attribution model?

A data-driven attribution model (or DDA model) is a default attribution model. This attribution model uses an algorithm to assign credit for conversions to various touchpoints in a conversion path.

In this model, values are assigned to touches in proportion to their contribution to conversion. The acquisition channel which assists the most gets the maximum credit for conversion and maximum resources are allocated to it regardless of it being a first touch, last touch or middle touch.

All other touches would get credit in proportion to their contribution to the conversion. For example, in the chart above, product reviews and product comparison sites played a very important role in my purchase decision.

So under the data driven attribution model, ‘product reviews’ and ‘product comparison site’ should get maximum credit for conversions regardless of them being the first, middle or last touches.

Similarly, had the clicks on the paid search ad helped me most in my purchase decision than ‘paid search’ would have got maximum credit for conversions even when it is neither the first touch nor last touch. All other touches would have got conversion credit in proportion to their contribution.

By default, the data driven attribution model is not enabled in a Google Analytics view, even when your GA property itself is enabled for GA premium.

So if you navigate to ‘Model Comparison Tool’  (under Conversions > Attribution) in your Google Analytics view:

Then click on ‘Select Model’ drop-down menu:

You won’t see ‘data-driven attribution’ model in the ‘Select Model‘ dialog box (by default):

To enable a data driven attribution model for your GA premium view, follow the steps below:

Step-1: Set up ecommerce and conversion tracking

Make sure that you have set up enhanced ecommerce tracking and/or goal conversion tracking in your GA view.

You need ecommerce and goal conversions data to generate a DDA model. Without these trackings set up, Google Analytics will not be able to generate a DDA model for you.

Setting up enhanced ecommerce tracking is not a mandatory requirement to generate a DDA model. You can also use standard ecommerce tracking.

The advantage of using enhanced ecommerce is that it provides much more ecommerce data than the standard ecommerce tracking which can be used by a DDA model to produce much better attribution model output.

Step-2: Link all of your Google Accounts to your Google Analytics account

In order for Google to generate the most accurate DDA model possible, you should link all of your Google Accounts (Doubleclick Campaign Manager, Google Adwords, Google Search Console, Google Play, Google Big Query, etc) to your Google Analytics account.

This is not a mandatory requirement for setting up DDA model but since the DDA model can analyse data from all Google accounts (which are linked to your GA account) and not just your GA account, so if you link all of your Google accounts to your GA account, then you would get exponentially better insight from your DDA model.

For example, if you use Doubleclick Campaign Manager (DCM) or Doubleclick Bid Manager (DBM) account then make sure it is linked to your Google Analytics account.

This integration will make you eligible for GDN impression reporting through which you can see view-through conversions data in your multi-channel funnel reports. 

Contact your GA premium/360 account manager and ask him to link your DCM/DBM account to your GA account and then enable the GDN impression reporting. Once the two accounts are linked and GDN impression reporting is enabled, then navigate to the ‘Admin’ section of your GA view, click on ‘All Products‘ link (under ‘Property’ column):

You should now be able to see DBM/DCM, under the ‘Linked products’ section along with the message ‘Receiving Data‘:

This confirms that your DBM/DCM account has been successfully linked to your GA premium property.

Do similar checks for other Google products and make sure all of your Google accounts are actively linked to your GA account.

Step-3: Integrate as much data as possible with Google Analytics.

This could mean, integrating your CRM, shopping cart, phone call tracking software, etc with Google Analytics.

This could also mean, setting up/fixing cross-domain tracking, cross-device tracking, and/or tracking offline conversions online in Google Analytics.

The more data you feed into your DDA model, from different marketing channels and devices, the higher will be the quality of your DDA model output.

Step-4: Import cost data into Google Analytics

In order to do ROI analysis in Google Analytics using a DDA model, you need cost data in your Google Analytics reports.

If your Google Adwords account is already linked to your GA account, you will automatically get cost data from Adwords in your GA reports. But for other paid marketing campaigns, you would need to import their cost data manually or via the management API.

Once you have uploaded the cost date for all of your paid marketing campaigns in your GA property, you will be able to measure ‘Data-driven CPA‘ and ‘Data-Driven ROAS‘ for each paid marketing channel:

Step-5: Meet the minimum conversion threshold for setting up a DDA model

Navigate to the GA premium view for which you want to enable data-driven attribution modelling.

Select that GA premium view that can meet the minimum conversion threshold for setting up a DDA model.

The minimum conversion threshold for setting up a DDA model is at least 400 conversions per conversion type with a conversion path length of 2+ interactions and at least 10,000 conversion paths in the selected reporting view in the last 28 days.

So your selected GA premium view must have recorded at least 400 Goal conversions and/or at least 400 transactions with a conversion path length of 2+ and at least 10,000 conversion paths, in the last 28 days period.

If your selected view does not meet the minimum conversion threshold requirement, then Google Analytics will not be able to generate a DDA model for you. In other words, the DDA model will appear in your selected GA view only once it meets the minimum conversion threshold requirement, for producing a DDA model. Otherwise, you may have to wait for weeks, to see a DDA model in your GA view.

However, if your selected reporting view, already meets the minimum conversion threshold requirement, then GA should produce the DDA model for you, within 7 days. So within a week, you should be able to see the DDA model in your selected view.

Step-6: Enable the data-driven model

#6.1 Navigate to the ‘view’ column (within the ‘Admin’ section of your selected GA premium view) and then click on the ‘view settings’ link:

#6.2 Scroll down the page and turn on ‘Enable Data-Driven Model’ (under ‘Modeling Settings’)

#6.3: Select DCM floodlight conversion type for which ‘data-driven attribution’ will be generated. You can select up to 20 DCM Floodlight conversion types. This step is optional if you do not use DCM.

Note: Double Click for Advertisers (or DFA) has been replaced with DCM by Google. You can select DCM Floodlight conversion types only when your DCM account, is linked to your Google Analytics account.

#6.4: Click on the ‘save’ button and then wait for a week, for GA to analyze your data and produce a DDA model.

#6.5: Navigate to ‘Model Comparison Tool’ (under Conversions > Attribution) in your DDA enabled GA view and then click on the ‘Select Model’ drop-down menu.

You should now be able to see your DDA model:

Issues with Traditional Multi-Touch Attribution Models

Let us consider that I followed the following conversion path:

From the conversion path above, we can conclude the following:

1. I read a blog post on your website.
2. After 3 days I saw your display ad on a website
3. After 2 days I read a review of your product on some website.
4. After 4 days I decided to make a purchase. So I made a search using a non branded keyword and clicked on your PPC ad on Google.
5. Just to make sure that I am going to get the best deal, I went to a product comparison site. Being satisfied with your product pricing I decided to make a purchase during the weekend.
6. During the weekend I again searched on Google but this time used a branded keyword and clicked on your organic search listing.
7. I made a purchase from your website.

In multi-touch attribution modelling, the conversion is attributed to multiple acquisition channels instead of just the first touch or last touch attribution. Here the middle touches also come into the picture.

This model aligns well the real-life situations as people rarely make a purchase through one or two acquisition channels. For e.g. it is highly unlikely for someone to read your blog post and then just make a purchase.

Similarly, it is highly unlikely for someone to see your display ad on a website and then directly make a purchase. He may read reviews of your product, go to a couple of product comparison sites before making a purchase. So we need to take all of the touches into account.

However, the problem with the typical multi-channel attribution model is the way credits for conversions are distributed to different marketing channels.

For example:

• Linear attribution model gives equal credit to each marketing channel/touchpoint in a conversion path.
• Position based attribution model gives more credit to first and last touch.
• Time decay attribution model gives more credit to the touchpoints which occur closest in the time to conversions.

In the real world, not all acquisition channels are equally valuable. 

For example, in the example above, I read product reviews and went to a product comparison site before making a purchase. These two touches were more valuable to me than the exposure to the blog post, display ad and the PPC ad as they play a very important role in my purchase decision.

Had I not been satisfied with the product review or pricing, I wouldn’t have made a purchase. Consequently, these touches should be given more credit.

Why is Data Driven Attribution Model better than traditional models?

It is better because of the following two reasons:

1. Data driven attribution model is a real-world attribution model.

That means it takes into account the back and forth activities of customers between multiple devices both online and offline while distributing credit for conversions. Because of that property, it has the ability to provide a much better picture of the conversion path followed by your customers. It is the first generation of truly multi-channel analytics modelling tools.

2.Data driven attribution model provides more flexibility than traditional models

The data driven attribution model takes into account your business model, marketing objectives, sales cycle, customers’ activities and seasonality as it allows you to assign credit to different marketing channels/touchpoints in proportion to their contribution to the conversion process.

Thus it provides more flexibility than linear, position-based and time decay models in terms of credit distribution.

Final Thoughts

One thing that you should not conclude from this article is that the traditional attribution models are flawed or useless.

This is because attribution is driven by experiments and in order to increase ROI across multiple marketing channels, you have to test different types of attribution models all the time.

It is only through continuous testing you can determine the acquisition channels which deserve maximum credit for conversions at a particular point in the time/product life cycle.

So it is critical that you do not sideline other attribution models (both single and multi-touch) in the favour of the data-driven attribution model.

Other articles on attribution modelling

1. How to analyse and report the true value of your SEO Campaign
2. How to valuate Display Advertising through Attribution Modelling
3. Understanding Shopping Carts for Analytics and Conversion Optimization
4. 6 Keys to Digital Success in Attribution Modelling
5. Google Analytics Attribution Modeling Tutorial
6. How to Measure and Improve the Quality of SEO Traffic through Google Analytics
7. How to explain attribution modelling to your clients
8. Default and Custom Attribution Models in Google Analytics
9. Understanding Missing Touchpoints in Attribution Modelling
10. What You Should Know about Historical Data in Web Analytics
11. Model Comparison Report Explained in Google Analytics Attribution
12. Data-Driven Attribution Model in Google Analytics – Tutorial
13. Conversion Lag Report Explained in Google Analytics Attribution
14. Selecting the Best Attribution Model for Inbound Marketing
15. How to do ROI Analysis in Google Analytics
16. Conversion Credit Models Guide – Google Analytics Attribution
17. Introduction to Nonline Analytics – True Multi Channel Analytics
18. Conversion Types Explained in Google Analytics Attribution
19. Attribution Channels Explained in Google Analytics Attribution
20. Differences Between Google Attribution & Multi-Channel Funnel Reports
21. Introduction to TV Attribution in Google Analytics Attribution 360
22. Conversion Credit Distribution for Attribution Models in Google Analytics
23. Conversion Paths Report Explained in Google Analytics Attribution
24. Attribution Model Comparison Tool in Google Analytics
25. Touchpoint Analysis in Google Analytics Attribution Modelling
26. Attributed Conversions & Attributed Revenue Explained in Google Attribution
27. Which Attribution Model to use in Google Analytics?
28. Google Attribution Access and User Permissions – Tutorial
29. Conversion Path Length Report Explained in Google Analytics Attribution
30. How to set up a data-driven attribution model in Google Analytics
31. View-Through Conversion Tracking in Google Analytics
32. Offline Conversion Tracking in Google Analytics – Tutorial
33. How to Create Custom Attribution Model in Google Analytics
34. 8 Google Analytics Conversions Segments You Must Use
35. You are doing Google Analytics all wrong. Here is why
36. How to Use ZMOT to Increase Conversions and Sales Exponentially
37. Connected Properties Explained in Google Analytics Attribution
38. Marketing Mix Modelling or Attribution Modelling. Which one is for you?
39. How is attribution modelling helpful for ecommerce and non-ecommerce websites?
40. Conversion Time & Interaction Time Explained in Google Analytics Attribution
41. How to Allocate Budgets in Multi Channel Marketing
42. How Does Attribution Work?
43. Data-Driven Attribution Model Explorer in Google Analytics
44. Introduction to Attribution Beta – Attribution Project in Google Analytics