Skip to main content
All CollectionsOptimizely
Optimizely - Setting Up a Web Experiment
Optimizely - Setting Up a Web Experiment

How to set up a Jacquard web experiment with Optimizely

Updated over a week ago

A one-time account configuration is required before you can get started with setting up an experiment. Check out our Getting Started article if you haven't already completed this setup.

Jacquard's Optimizely integration workflow

Jacquard's integration uses the below workflow to automatically push your optimised Jacquard language variants directly into your Optimizely Experiments and feedback your results into Jacquard for ongoing optimisation.

Set up a Jacquard experiment in Optimizely

1. Log in to your Optimizely account and click Create New... and then A/B Test within the project set up for Jacquard tests.

2. Configure the settings and URL for the test and click Create Experiment.

3. Delete Variation #1 from the list of variations. Jacquard will automatically create your language variations for the Experiment in a later step.

Click Edit next to the Original variation. This variation will be the human control variant you would display on your site if you weren't using Jacquard.

4. In the Optimizely editor, select the element on the page that you want to test. Update the element with your human control copy for the test. Click Save.

4. Open your Jacquard account and select Create a new experiment. Name your experiment and select Web & App as the campaign type. Enter your human control language and choose the number of variants you want to test.

Generally, we recommend 5 variants for a web test, but choose the number of variants discussed with our Data Science team during your methodology consultation. If you have not yet had one, reach out to your Customer Success representative to schedule one.

Click Next.

5. In the Language Generation step, select the start date of your experiment under Send date and then the use case for the experiment you're running (i.e. the language element that you are testing). Click Generate to create your language variants for testing.

6. Review the language variants. You can edit the variations by hovering over the variant and selecting Request a change on the right side of the screen.

For best results, avoid editing the Jacquard variants unless something is incorrect or needs updating for reasons such as industry compliance or your brand's guidelines.

Click Approve and continue.

7. In the panel on the right side of your screen, click the Optimizely integration button to open the integration drawer.

8. Within the integration drawer, select the Optimizely Project that contains your new A/B Test. Next, select the Optimizely Experiment you just created. Click Create and Send to publish the variants into your A/B Test for your review.

If for any reason you need to unlink campaigns, click Unlink.

Clicking Unlink will not delete the variations from Optimizely. You will need to do that manually after unlinking the Jacquard experiment.

9. Navigate to your Optimizely account and refresh the page to see the language variations you generated in Jacquard. Preview and verify the variants using the before finalising test setup.

10. Select your Audience, Metrics, Traffic Allocation, and Schedule for your test.

Audience: Configure your audience for the test. This is set to all website visitors by default, but you can create different Audience Segments as required. Click Save.

Metrics: The metrics available here are based on the Events configured at the project level before setting up the experiment. Select your Primary Metric for the Experiment and click Save to Experiment. Click Save.

You can select additional Events to track in Optimizely, but the Primary Metric selected is what will be pulled back into Jacquard for ongoing optimisation.

Traffic Allocation: Select either Manual or Stats Accelerator for your Distribution Mode. Click Save.

Select Manual if you would like to manually set the traffic distribution for each variation in the experiment.

Select Stats Accelerator if you would like to manage the audience distribution dynamically.

Schedule: Enter the timezone, start time, and end time for your experiment. Click Save.

11. Publish your Experiment by selecting Start Experiment and the option Publish Only.

12. Navigate back to the campaign you set up in your Jacquard account and open the integration drawer. Select the Primary goal and Data collection schedule to match what you selected within your Optimizely experiment.

Jacquard will not start the experiment in Optimizely. The timeframe selected here informs Jacquard when to start and stop collecting results for the campaign. You will still need to start your experiment in Optimizely when ready.

13. Click Save Schedule. Once your Experiment begins, Jacquard will pull results in for each variant on a nightly basis for the duration of your experiment.

The option to stop collecting data will appear once you have saved the schedule. This allows you to end the collection of results or update the schedule as required. To do this, simply click the Stop collecting data button in the integration drawer.

Did this answer your question?