Lesson 1: Introduction to Attribution Flashcards
What is attribution?
Attribution is the practice of tracking and assigning values to all marketing touchpoints that lead to a conversion
What is a touchpoint?
A touchpoint is an online interaction or event. It could be an impression or a click type interaction in a conversion path.
What is a conversion?
A conversion can be a website purchase, an email sign up, whatever is important to the business
What is an attributed conversion?
An attributed conversion is a conversion event that can be credited to multiple impression or click events
What are channels?
The set of media channels that make up your media investment portfolio, defined specifically by client
What is unique about the data-driven model?
The data-driven model does not have any pre-determination. Credit is assigned after the model has been run.
How does the data-driven model work?
The data-driven model uses a machine learning algorithm to assign a weighted fraction to each marketing touchpoint based on the touchpoint’s influence on conversion.
How does the machine learning algorithm work?
It is an A/B test approach that compares paths and takes into account factors like exposure, recency and frequency.
How does the machine learning algorithm compare conversion paths?
It will compare like paths and compare the probability of conversion between path A and path B when one of the touchpoints is missing.
How is the weight of each touchpoint assigned?
The weights given are proportional to the relative influence of each touchpoint in driving a conversion.
How is the credit divided between each touchpoint?
Credit is fractionally divided up to all impressions and clicks on the user path.
What are the key factors used to determine the credit assigned to each touchpoint?
Lookback Exposure Recency Frequency Frequency at each value of recency METs Anonymity Cross Device model
What is lookback?
Lookback is the max amount of time allowed between the conversion and the touchpoints
What is exposure?
Exposure defines the event type the user was exposed to . It can be impression or click event.
What is recency?
The time in days between each event time impression or click and the conversion or estimated customer drop off
What is frequency?
The count of events impressions or clicks for each type of event in the path.
What is an example of frequency at each value of recency?
User saw 2 impressions, 3 weeks ago from touchpoint x (paid search).
What can you see with the cross device model?
You can see first interaction and last interaction across device.
What provides the data for the cross device model?
The Google Device Graph powers cross device attribution
What is the Google Device Graph
The Google Device Graph stitches together path fragments for users. It ties device islands together based on users behaviors on google biggest online products.
How can the cross-device model impact reports in Attribution?
It can change the number of converting and non-converting paths by stitching together what might seem like 3 paths into one.
When you pick a date range in Attribution reporting, that is the date range associated with the (blank)
the media
All of the (blank) have to take place in the data range to be counted in the reported period
touchpoints
In attribution reporting, early funnel touchpoints launched today will take (blank) to see conversion
time
In attribution reporting, conversions will continue to accumulate (blank)
daily
In attribution reporting, credit is no longer assigned the same day as the (blank)
conversion
In attribution reporting, the longer we wait to look at data after the reporting period ends, the more conversion credit will be attributed, particularly to (blank) funnel touchpoints.
upper funnel
In attribution reporting, conversion numbers will have settled by the end of the (blank) period
lookback
Lower funnel media tends to get attribution credit faster than (blank)
Upper funnel
Conversions stop accumulating in attribution reports after (blank)
The Lookback period ends
In attribution reporting, the period of time when attribution conversions continue to accumulate is called the (blank)
the dynamic window
The key metric you can analyze with Attribution is (blank)
how many conversions did I get for what I spent on media? cost per attributed conversion
What is exposure in Google Attribution
Exposure is what a user has been exposed to, for example the type of event like paid search
What is a normalized score?
A normalised score is a score that has been scaled up so the total values can equal 100%
What does MET stand for in attribution?
MET stands for model event type, the type of event a user was exposed to
Each model event type and recency bucket combination is treated as a (blank) item when calculating recency
Each model event type and recency bucket is counted as a separate item
When calculating the impact of an event, PPC generic at recency 5-7 days is a (blank)
PPC generic is a model event type in one recency bucket
Recency tells us how many days prior to conversion the (blank) occurred
Recency tell us how many days prior to conversion the event occurred
The counterfactual represents the path to conversion with one event (blank)
removed
