The Did [not] Perform clause allows you to limit query rows to users who did or did not do a selected event before or after performing the initial event. It is a great way to examine users who performed more than one event in a sequence you can define in Indicative. Here's a simple example: you could use a did [not] perform clause to view users who did "Subscribe" after "Email Clicked".
Available in: Segmentation
- A query row that already has an initial event
"Did [Not] Perform" Clause Basics
Creating a "Did [Not] Perform" Clause
To use a Did [not] Perform clause, select an event from the Did [not] Perform data dropdown. Continue through the rest of this article for modifications that will provide you the flexibility to understanding your users.
Did vs Did Not Do
After selecting an event in the Did [not] Perform dropdown, decide whether you want to understand the data of users who did event or did not do this event. For example:
- Show the total count of events where users who did Subscribe also did Email Clicked at least once within the prior 7 days.
- Show the total count of events where users who did Subscribe did not do Email Clicked within the prior 7 days.
Note: Switch from Total count of to Users who performed to examine the count of users rather than the count of events but keep in mind that this is the dropdown that determines whether the query returns an event count or user count.
Adjusting Event Count
After adding an event using the Did [not] Perform clause, you can adjust for the number of times users performed this event, such as greater than or equal to 1 time, less than 5 times, or greater than 3 times. For example, you can view the event count for users who did Subscribe, and also did Email Clicked greater than 3 times within the 7 days prior to subscribing.
Adjusting Date Range
After adjusting the event count, customize the date range. The first menu allows you to select whether the Did [not] Perform clause event should have occurred prior to the initial event, subsequent to the initial event, or between two selected calendar dates. When using between, the second menu displays a calendar, allowing you to select a specific date range.
When using within the prior/subsequent, the second menu allows you to select a preset or custom count based on minutes, hours, days, weeks, or months.
Multiple "For" Clauses
Multiple Did [not] Perform clauses may be added by selecting additional events from the Did [not] Perform dropdown.
As events are added, you can adjust the event count and date range of each additional Did [not] Perform clause. When using a relative date range, the date range is always relative to the date of the base event in the query and not the other Did [not] Perform clauses.
For example, you could view the event count for users who did Subscribe for users who also did Email Clicked 2 times or more within the prior 7 days, and also did Blog View 3 times or more within the prior 7 days. This will show users who clicked email links and viewed content within 7 days before subscribing.
Difference between Funnel and a Did [not] Perform Clause
The most common method of analyzing a user journey is to use the Funnel tool. The Funnel tool is specifically designed to measure user journeys, and is highly customizable through features such as Conversion Precision, Optional Steps, and Tracking Properties. However, user journeys combined with a Conversion Limit can be recreated in a Segmentation query combined with a did [not] perform clause. However, these two methods can yield different results. Consider the following queries:
Did [not] Perform:
These two queries are measuring the same user journey - a user who does Email Clicked, and then within 24 hours also does Subscribe. However, the funnel query shows a considerably lower amount of converted users. This is due to the following:
- A Funnel query only counts a user once. Therefore, the funnel tool will only consider the first instance of an event.
- A Segmentation query will count users more than once. Therefore, the segmentation tool considers all instances of an event.
Consider the following example:
- User performs Email Clicked at 10:00 AM on 2/27
- User performs Email Clicked at 13:00 PM on 2/27
- User performs Subscribe at 11:00 AM on 2/28
The funnel tool will look at the first instance of Email Clicked (at 10AM), and see that the user did not complete Subscribe within 24 hours of that instance of Email Clicked. This user is therefore not counted in the funnel analysis.
The segmentation tool will look at all instances of Email Clicked, and therefore see that the user performed Email Clicked at 13PM, and Subscribe within 24 hours of that event. This user is therefore counted in the segmentation analysis.
So, when comparing user journeys in Funnel to user journeys in Segmentation, the user count will always be lower in Funnel due to the fact that the Funnel tool only counts each user once, and looks at the first instance of the initial event.
How It's Done
Take a look at the following example query:
To decide which blog view events will be counted, Indicative will look at the timestamp on each blog view and check to see if there is a create profile event that falls in the 7 subsequent calendar days. If there is, the blog view is counted, otherwise not.
Keep In Mind
Note that the Did [not] Perform clause events could have occurred outside of the main date range in the query builder.
For example, Subscribe events in the chart window could have occurred within the Last 7 Days of the current date, as set in the date range menu at the base of the query builder. However, the Email Clicked and Blog View events will have occurred within the 7 days prior to the Subscribe event.
If a Subscribe event occurred 6 days ago, and Email Clicked occurred 6 days prior to that, the Subscribe event will be counted even though Email Clicked occurred 12 days prior to today.