Building an Impact Analysis

An Impact analysis is built by measuring retention or engagement using the specified conditions against the segments of actors and attributes for a given time frame. It allows users to measure the impact of an event on user engagement. The results of an impact analysis are anchored around the timing of the impact event, and the measures aggregate the impact across users.

As an example, let's generate an impact analysis to assess how much the Make Purchase action influences users to view similar products on the website.

1. Access the Explorations page on the left navigation panel.
2. Select the Impact template.
3. On the definition page, choose the preferred measure. Here, we choose Average Events Per Actor. This will measure the average number of events per actor that performed any matching activity during the selected time bucket. The average is computed regardless of segmentation being used. Also, choose the Users dataset.
4. In the Events and Outcomes section, define the treatments and outcomes - that is, the events that we want to measure the impact for. For the example, let us say that we want to see the impact of the Make Purchase treatment on the View Similar outcome. If you want to use a Timestamp instead of an Event, you can click the Timestamp button in this section.
5. Then, choose Event Occurence. There are 3 options that you can choose from - for the first time in the time range, for the first time ever, or for the last time. For our example, we set it to for the first time in the time range.
6. As the next step, we define Bounds - the time before and after the first occurence of the outcomes defined in the previous step. Here, we set both to 30 days.
7. Set the time grain to 30 days.
8. On the Visualization pane, set the time range and time grain. Click Run Exploration to generate an Impact chart.
9. Give a descriptive name to the analysis and save it.