In marketing campaigns, marketers send campaign ads to selected customers ("test group") and putting aside the randomly selected some group of customers ("control group") from the campaign. The idea is to determine the effectiveness of the compaign by comparing the sales generated from the "test group" and "control group".
For example, an ecommerce website has segmented customers to send "50% discount on clothing" campaign and they send this campaign to only 50k of them (“test group”), put aside a randomly-selected 5k customers (“control group”) who will not receive it. Once the campaign is over, they will determine the effectiveness of the campaign by comparing the additional revenues generated by the test group with those generated by the control group.
The following table provides a comparison between two groups of customers - the "Test" group and the "Control" group. In marketing, the 'Test' group refers to the target audience (users) that receives email or phone call campaigns as part of a new advertising strategy or promotional offer. Whereas, the 'Control' group refers to users who are not part of the campaign or promotional offer.
| Variables | Test | Control |
|---|---|---|
| Customer Size | 50k | 5k |
| Purchase Rate | 15% | 10% |
| Avg Spend Per Customer | $150 | $125 |
| Total Sales | $1,125 | $63 |
 |
| Campaign Result : Test vs. Control |
In this scenario, we assume that the company has not sent any campaign or promotional offer to any user.
| Variables | Without Campaign |
|---|---|
| Customer Size | 55k |
| Purchase Rate | 10% |
| Avg Spend Per Customer | $125 |
| Total Sales | $687.5 |
 |
| Test vs. Control Methodology |
The total sales generated as a result of the campaign would be the sum of the sales from both the test and control groups, while the profit generated from the campaign would be calculated by subtracting the sales without the campaign from the total sales.
Total Revenue from Campaign = $1,125 + $63 = $1,187.5
Profit from Campaign = $1,187.5 - $687.5 = $500
Deepanshu founded ListenData with a simple objective - Make analytics easy to understand and follow. He has over 10 years of experience in data science. During his tenure, he worked with global clients in various domains like Banking, Insurance, Private Equity, Telecom and HR.
Really useful!! Thank you!
ReplyDeletea
ReplyDeleteThank you :)
ReplyDeleteWhat is this term called ??
ReplyDeleteAre their other methods to create the control group where it is not random and the control groups "looks like" the test
ReplyDeleteyou can take out stratified samples for control making sure that it has the same feature distribution as compared to the test group
Delete