👉Configuring a Progressive rollout

📖 Definition

A Progressive rollout (or Progressive deployment) enables you to progressively release a feature to your users by choosing each deployment step and the proportion of traffic allocated to your users.

⚙️ Configuration

To configure a Progressive rollout, apply the following steps:

1. From the dashboard, click Create a use case.

2. Select the Progressive rollout template. [Basic information]

3. Enter the name of your feature and its description.

4. Choose the primary and secondary KPIs you want to follow.

5. Click Save and continue. [Scenario]

6. In the next step, define your targeted users for your progressive rollout feature & the flag's name, type (text, number, boolean, array, or object), and value that will control the deployment of your feature.

7. Click Save and continue. [Steps]

8. Then, configure each step of your feature deployment. You can choose between a classic and a statistical rollout. In the classic rollout, visitors are either assigned to a variation or untracked. No traffic is assigned to the original version, so no comparison is possible between the original version and the variation. In the statistical rollout, you can define deployment steps and configure a rollback KPI which enables you to define, for a specific event, a percentage that you don’t want to exceed from a specific number of visitors. When using this option, you avoid any type of product regression due to the deployment of your feature. For more information on the rollback KPI, refer to Using the automatic rollback option.

9. Click Save and continue. [Overview]

10. Check that every step has been configured correctly.

11. (Optional) Notify your teammates that progressive deployment is ready to go.

⭐ Good to know For each type of rollout, you must define the date, time, zone and % of users involved in the first wave of deployment. Then, you will have to choose between a deployment with fixed steps or customized steps by defining the % of users allocated and the frequency of each deployment step.

💡 Use case

Let’s say you want to develop a new feature, allowing your users to know the exact location of their package. To mitigate the risk, you want to deploy this feature to your users progressively. In addition, to avoid product regression, you want to keep your conversion rate above 8.2%.

To do so:

1. Create a Progressive rollout use case on your Flagship account.

2. Choose the KPI you want to follow. Here, you can select the conversion rate generated by your feature

3. Configure the flag that will control your new feature.

4. Set your targeted users to All users so that all your visitors can see the feature.

5. Configure your deployment steps as a statistical rollout and with fixed steps of 10% of the users every week until you have reached 100%.

6. Configure your rollback KPI: Select Event as the KPI type and the name of the conversion rate measurement KPI you have chosen for your feature.

7. Select the < operator and enter the rate (8.2%).

8. Enter the number of users from which the rollback can be done in case the conversion rate is lower than 8.2%.

9. Check the overview of your Progressive rollout.

10. Save and activate your use case via the dashboard.

Need additional information?

Submit your request at [e-mail]

Always happy to help!