For years, AdWords has been operating on a Last Click Attribution Model. This means that only the last Adwords Ad that was interacted with before a conversion was given credit for that conversion. This allows for advertisers to know exactly which Campaign, Ad Group, Keywords, Audience, etc. led directly to a conversion.
Although the Last Click Attribution Model certainly has value, it doesn’t show the customer’s entire funnel. It ignores all the upper funnel touch points that ultimately lead to a conversion. For example, let’s say a customer is looking to buy a gift for a friend. They may start their research by looking up “tech gifts for friends”. The next time they go to search, they look for the “highest rated virtual reality headsets” and then finally search “Google Daydream View”. In this scenario with Last Click, only the “Google Daydream View” keyword receives credit for the conversion. This can cause advertisers to severely undervalue the importance of their upper funnel search campaigns in the account.
To keep up with the increasing complexity of a user’s buying experience, AdWords now offers advertisers the ability to select an alternative to the Last Click Attribution Model in Adwords that distributes credit across search ad clicks. There are 5 models to choose from (these models are only available for search campaigns):
The Data Driven Model is one of the newest models offered within the AdWords interface. This model uses your account data and algorithms to determine the impact each keyword had on a user’s likelihood to convert. This model uses factors such as ad interactions, order of exposure, and the creative assets used to figure out which interactions were the most critical in driving conversions. It then allocates conversion credit based on each keyword’s contribution to conversions. Since this model uses such advanced algorithms custom to your account, it is recommended that you use this attribution model if you’re account is eligible. In order to be eligible to use the Data Driven Model, your account must have at least 20,000 clicks and at least 800 conversions within the 30 days. Learn more about the Data Drive model here.
See some benefits of switching to data driven attribution below:
- Values all steps on the conversion path
- Works with automated bidding like Target CPA (as do the other attribution model options)
- Works even on very short conversion paths
- Quick and easy to implement
- Regularly updated based on changes in account
- Uses 100% of the accounts data to learn over time
If you’re not eligible to use the Data Driven Model, it is recommended that you use a Rules Based model such as Time, Decay, Linear, or Position Based. Ruled Based models use fixed rules to assign credit to touch points. While Rule Based models don’t use advanced algorithms, they still help us understand how our marketing channels work together beyond the Last Click model
See below some best practices when choosing an attribution model:
- Before making changes, compare conversion and cost data of different attribution models to understand how conversions would shift using the AdWords Attribution Modeling Tool
- Use the Path Length metric to see how many clicks on average happened before a conversion. The longer the path, the more your account will benefit from switching from a Last Click model. See how to get to the Paths report here.
- See our Attribution Best Practice Guide, “Beyond Last- Click Attribution”, here
These new multi-touch attribution models now give us the ability to “connect the dots” and understand the whole conversion path within your campaigns. It allows us to understand our customer’s behavior starting from awareness to decision across multiple devices and channels. Armed with more information, these models allow advertisers to better leverage the data within their account to make more educated marketing decisions.
Already made the switch? Let us know about your experience with a Multi Touch Model!
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.