Process: Create Marketing Experiment
Flow
Objective
The creation of a marketing experiment is the initial step in the overall experimentation process. The marketing experiment represents an alternate version of site content, whose effectiveness will be compared against the original, or control, version.
Description
In order to measure the effectiveness of a marketing campaign, the Marketing Manager will create a marketing experiment. In the experiment, the Marketing Manager selects a content that represents the current content shown in the site, known as the control element. The Marketing Manager then changes a factor of that content for testing. The types of factors that can be tested are:
- content - content displayed in an e-Marketing spot generated by the web activity
- segment - customer group that the web activity targets
- spot - e-Marketing spot in which the content generated by the web activity is displayed
- activity - the Marketing Manager can choose to test the entire web activity by swapping the control with another pre-defined web activity instead of only swapping one of the factors,
At least one element of the type of selected component is selected as 'experiment', which is known as test element. Each test element represents a changed version of the Web site or an alternative site content. Customers are selected to be shown either the control element or a test element according to a percentage split defined for each element. Information about what was shown, when and to whom is captured for analysis. At the end of the experiment, the winning element is made permanent in order to maximize the effectiveness of the content.
Edition
Professional, Enterprise
Tasks
| Task | Description | Role |
|---|---|---|
| Select control element |
The creator of this marketing experiment selects a control Content. |
Marketing Manager |
| Select test elements |
The creator of this marketing experiment selects test elements, which can be Web activities, displaying contents, customer segments and e-Marketing Spots depending on the experiment type defined during the design of this experiment. |
Marketing Manager |
| Set experiment split ratio |
The creator of this marketing experiment sets a ratio for the control element and each test element. The ratio number determines the percentage of unique customer visits for which the test or control content will be displayed in the e-Marketing Spot. |
Marketing Manager |
| Set up marketing experiment properties |
The creator of this marketing experiment sets the name, description, priority, persistence scope which determines how long the same experimental content will be shown to customers before a new one is selected, start date, and expiry options including end date and the number of customers who viewed the experiment before it expires. The creator also needs to decide the type of experiment he/she would like to run. Here are the four available types:
|
Marketing Manager |
Business artifacts
(C) Copyright IBM Corporation 1996, 2006. All Rights Reserved.