Multi-Touch Attribution - Overview

Overview

Attribution analyzes a user's marketing interactions as he or she moves towards a target event (like a website conversion). The goal is to determine which interaction - or combination of interactions - has the biggest influence on their decision to perform a target objective.

The Multi-Touch Attribution dashboard uses Google Analytics interactions (sessions) to record all of the user's interactions prior to completing a target event on the website. The attribution values are then combined with OneView Media performance data to calculate ROI for each model. 

The dashboard compares the results of many different attribution models to allow users to see the difference in their data. The models are:

• First
• Last
• Linear
• U-Shaped
• Smooth-U
• Bell
• Time Decay
• Data-Driven

Below are links to Help Center articles explaining the models:

• Rule-Based Attribution Models
• Data-Driven Attribution Models

Dashboards

This report includes the following dashboards:

Attribution Analysis:

1. Is our current reporting strategy/attribution model the right method for our business?
2. What is my ROI based on my reporting strategy?
3. How much does each of my marketing channels contribute to business objectives?

Model Comparison:

1. What are the differences in an interaction type's impact between different models and methodologies?

Next Interaction:

1. How well do my interaction types work with each other? What is the probability that one interaction will lead to another?

Top Paths:

1. What are my most common path groups?

Supported BI Tools

The Multi-Touch Attribution Dashboard is available in Tableau and Power BI.

Data

This dashboard includes the following data sources:

• Google Analytics
• OneView Media performance

This dashboard is built on the following datasets:

• v_cmx_attribution
• v_cmx_top_paths
• v_cmx_transition_matrix

Video

FAQs

Below are commonly asked questions about the Multi-Touch Attribution Dashboard:

1. What do these date windows mean? What is the difference between the Target Date Window and the Lookback Window?
   • These date windows allow you to customize what data is included in your model. While the Target Date Window is focused on what paths are included in reporting, the Lookback Window is focused on which interactions within a path should be included. Please read our article on ChannelMix Modeling Engine's Date Windows to see in-depth explanations of how these windows work and how they work together.
2. What does "Filter Direct" mean? I don't want to remove Direct traffic, why would I use this? What happens when I select yes to "Compress Paths"? How do these options affect my reporting?
   • "Filter Direct" and "Compress Paths" are configuration options that can be used to customize the way each user's path is constructed. Please read our article on Path Configurations for explanations on each option and when to use them. 
3. There are a lot of different options, how do I choose which model is best for me?
   • Selecting your attribution model depends on your business. Do you know what you are trying to measure? How well do you know your customer's paths? Check out this article for some tips on how to choose the right attribution model for you. 
4. This dashboard is meant to combine my performance from my Media OneView. Why don't all of my channels appear in the Channel Performance section of the dashboard?
   • Only channels that appeared in paths (per the lookback window) will be assigned an attribution value. There may be scenarios where you have spent money in a channel but an interaction did not occur during the date windows you have set. A common occurrence is when you introduce a new channel to your marketing. Viewers of these ads may not have interacted yet, or have had the chance to convert on your website yet. 
5. I see the option for a Data-Driven model in my dashboard, but it is blank. Is there something wrong?
   • Adding the Data-Driven option to the Multi-Touch Attribution dashboard requires first-party functional cookie implementation - including ChannelMix Conversion Tracking and ChannelMix ID. If you are interested in this, please read our article on our Data-Driven Attribution and reach out to your team at Alight Analytics.
6. Not only is the Data-Driven option blank, but the entire dashboard is blank. Is there something wrong?
   • A blank dashboard indicates an incomplete analysis. This would be due to your target event not occurring during the target date window. If your dashboard is blank, please reach out to your team at ChannelMix to discuss business strategy and objectives.
7. I am looking at the Interaction Heat Map in the Next Interaction tab. What does Initial mean? What about Next, what does End mean?
   • The Interaction Heat Map is meant to show the probability of one interaction leading to another. That said, Initial represents an interaction you are looking at first, while Next represents the following interaction. This can be best understood in a sentence, for example, the probability that an interaction in Paid Social (Initial) will lead to an interaction in Paid Search (Next) is 50%.
   • End represents the end of your path - which would be the completion of your target. If your interaction type has a high probability of ending, that means that interaction type is great at closing your customer's path. 
8. What are the groups in my Top Paths? I know I have paths where people come to my website through Paid Social multiple times. How come I do not see a group that shows Paid Social more than once?
   • The main idea in the Top Paths tab of the dashboard is to provide which interaction types are appearing together in customer paths. Here we do not consider how many times each interaction type appears in a path, we only consider what interaction types appear in a path. For a more detailed explanation of this, please check out this article on Understanding Top Paths.