1. Which Attribution models can be used in the Data Studio? How do they work?
Admetrics Data Studio offers seven attribution models that you can easily switch between and compare. There’s no “right” or “wrong” attribution model, each has its benefits and drawbacks. Reviewing the output of different models helps you understand the contribution of each channel and campaign, guiding your decision-making.
Our attribution models are based on the user journeys tracked via our pixel integration. Below is a breakdown of each available attribution model:
First touch attribution
The first touchpoint within the attribution window that is not direct receives 100% of the orders and revenue.
Last touch attribution
The last touchpoint within the attribution window that is not direct and not organic gets 100% of orders and revenue attributed.
Last platform touch attribution
The last touchpoint of each paid platform receives 100% of the orders and revenue. This model inflates and duplicates orders across platforms, while the Direct channel is ignored. This approach helps compare performance across platforms but is not recommended for cross-channel evaluation.
Tip: For cross-channel evaluation, use a fractional multi-touch model, such as Linear or Position-based Attribution.
Linear attribution
Each non-direct, non-branded touchpoint within the attribution window gets an equal share of the orders and revenue. For example, if there are 3 touchpoints, each will receive 33.3% of the attribution.
Tip: Paid brand search channels are mostly excluded from this model, configured via the channel settings.
Linear-direct
The Linear-Direct Attribution Model distributes credit for a conversion equally among all touchpoints, excluding direct visits (when users visit your site by typing the URL directly). This model ensures that each non-direct interaction in the customer journey, whether it be through paid search, social media, or organic traffic, receives an equal share of the attribution. It is ideal for businesses looking to understand the value of all marketing channels involved in driving conversions, except for direct traffic, which is not considered an active marketing effort.
Position-based attribution
The first and last touchpoints (that are not direct and not brand search) receive 40% of the orders and revenue each, while the remaining touchpoints share the remaining 20%.
Paid brand search channels are mostly excluded from attribution in this model. This is configured through the Channel group setting in the channel configuration (”Paid Branded Search”).
Inflated attribution
Every touchpoint within the attribution window gets 100% of the orders and revenue, resulting in inflated and duplicated attributions across platforms. This model is useful for understanding relative performance within a single platform.
Tip: Use fractional models like Linear or Position-based Attribution for cross-channel evaluations.
Any touchpoint within the attribution window gets 100% of the orders and revenue attributed. This model will inflate and duplicate orders even within platforms.
2. What is a lookback window?
A Lookback Window is the number of days prior to a conversion when touchpoints are considered for attribution. You can customize your Lookback Window in the Data Studio, with options ranging from 1 day to "Max" (infinite). Here it is explained how you can choose your Lookback window in the Data Studio.
3. How long is a user session?
A user session lasts for a maximum of 30 minutes. If no new events are logged within this timeframe, a new session is created. Each session results in a visit.
4. Why shouldn’t I rely on media partner self-attribution?
Media partners often have a limited view of the customer journey, meaning their attribution models might not be fully reliable. Here's a chart that visualizes the narrow perspective of media partners:
5. Summarized information in the video:
This video explains:
why attribution is important
how Admetrics does tracking with the help of UTM-parameters
what attribution models exist
the perks of each model and when you should use which model
why the attribution of media partners are often not correct