Which data is attributable?
Webmarketer works out how your marketing actions contribute to three types of items:
- Events
- States
- Statistics
Events describe an action performed by the user, such as a purchase, a sign-up, an add-to-cart, and so on.
States and statistics, on the other hand, are pieces of information describing those very events: they capture the progress of an event over time, or numeric data attached to it (signature, revenue, and so on).
Events
Events are sent to Webmarketer when a user performs an action on your website. When attributing, Webmarketer looks through every interaction that happened within the attribution window configured on each attribution model.
Any interaction that happened between {event date} - {attribution window duration}
and {event date} is considered during attribution.
For example:
- If an event happens on February 10 and the attribution window is 7 days, Webmarketer considers every interaction that happened between February 3 and February 10.
- If an event happens on February 10 and the attribution window is 30 days, Webmarketer considers every interaction that happened between January 10 and February 10.
Note that there is no need to attribute every single event happening on your website to marketing campaigns; we generally recommend sticking to events that carry economic or strategic value for your business (purchases, quotation requests, sign-ups, add-to-cart).
States
Events can be complemented by states, which are boolean values (true/false) describing the progress of an event over time.
For example, a "Quotation request" event can carry the states:
- "Quotation sent", turned "true" on March 17
- "Quotation signed", turned "true" on March 20
- "Quotation cancelled", turned "true" on March 25
This makes it possible to measure the progression of an event over time and to identify the marketing campaigns responsible for each of those states.
The date used to attribute a state is the date the state changed, and not the date of the event.
That distinction matters, because it accounts for any new ads seen by the user in the meantime, which may be responsible for the move to a new state.
Any interaction that happened between {change date} - {attribution window duration}
and {change date} is considered during attribution.
Statistics
Like states, statistics are evolving pieces of information attached to an event. They describe numeric data (revenue, quotation value, number of products, and so on).
As with states, each statistic can be set on an event at a specific date.
Typically, a "Quotation request" event will carry the following statistics:
- "Quotation value", €1,000 on March 17
- "Number of products", 5 on March 17
- "Revenue", €800 on March 20 (when the quotation was signed, after a discount for instance)
The date used to identify the contributors is the date the statistic changed, and not the date of the event.
As with states, that distinction matters, because it accounts for any new ads seen by the user in the meantime, which may be responsible for the change to the statistic.
Any interaction that happened between {change date} - {attribution window duration}
and {change date} is considered during attribution.