Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 4 Next »

Isolate CRM incremental impact on total sales from loyalty incrementality by relying on the control group methodology

What are we trying to measure?

Through this methodology, we want to measure the impact of Loyalty & CRM over total sales. On top of that, we want to isolate the impact of CRM from the one of loyalty

Methodology

Create 2 control groups:

  • Control group for opted in people, to measure CRM incrementality over total sales

  • Control group at campaign level, to assess efficacy of individual campaign & incentives and optimize the CRM strategy

How to set up the groups

1, Open your Braze account and click on Segments

2. Select from the drop down list Custom Attributes

3. Create a new segment with the following characteristics:

  • MarketingEmail = True OR

  • MarketingPush = True

By relying on the OR logic, we are selecting all guests who have opted in to at least one channel (either email or push, or both)

4. Export this cohort and randomly select 10% of the users. This 10% will be the fixed control group

5. Import the file in Braze by clicking on create a segment and selecting the CSV file

6. Include the cohort as a control group for all campaigns, both automated campaigns & journeys. To exclude the control group from the campaigns, in the Target Audience select Consumer Membership and Is Not part of the control group

7. Export all the cohorts (opted out, opted in, control) in Amplitude by:

  • Clicking on User Data and then CSV Export Email Addresses

  • In the export, only keep the User ID column

  • Open Amplitude and click on Create New

  • Click on User Cohorts

  • Select import from cohort file

  • Select User ID

  • Create the cohort and when you are done click on Save

  • Give your cohort a title and click on Save

8. Create a new analysis, selecting segmentation and use as event Online Loyalty Transaction Claimed & In Store Loyalty Transaction Claimed all together (all RS). You can adjust the timeframe where the condition have to be verified, and the period you want to look at.

By selecting “daily” the unique users will be counted each day. By selecting weekly, the unique users will be determined on a weekly basis. Therefore the sum of the weekly unique does not correspond to the monthly unique

9. Include as segmentation criteria where cohort = test group and control group

10. Extract unique, event totals and monetary value (note: to extract monetary value, select group by subtotal and then click on properties, and select sum of property value)

11. Calculate frequency, check and monetary value per user by using as denominator the fixed cohort

12. Calculate the pre-post NoC by comparing the behavior of the test group with the behavior of the control group. Eg:

Fixed cohorts

  • Test: 10k

  • Control: 1K

Unique per week

Week 1

Week 2

Week 3

Week 4

Test

120

121

130

133

Control

12

11

11

12

Events Total

Week 1

Week 2

Week 3

Week 4

Test

140

141

160

173

Control

12

12

11

14

Monetary Value

Week 1

Week 2

Week 3

Week 4

Test

800

810

780

790

Control

80

81

79

80

Frequency = events total by week by cohort / fixed cohort size

Check = total sales by week by cohort / total events by week by cohort

Monetary value = frequency x check

  • No labels