Checklist before Deployment

Seems like you are almost there! Here are some things you should double check before deploying.

๐Ÿ‘

Make sure you have Tuned LiftIgniter correctly

Make sure that the recommendations returned are what you want, confirm that Widgets are rendered and tracked correctly, and anything you don't want to show is hidden.

You should be using an A/B test! If you are comparing to an existing recommendation widget, you'll A/B test against that area. If you are launching a new recommendation widget, you can A/B test against nothing. Both cases are included in our documentation on Full Integration with A/B Testing.

In order to make sure that the tracking events are firing correctly, you can set _debug to true and check for alerts for visible and click events. For more, see our Tracking Widgets documentation.

๐Ÿ‘

Don't forget the console logs

Before deploying to production, make sure to remove all console.log statements and any statement setting _debug to true.

๐Ÿ“˜

Use runDiagnostics

Call our $p("runDiagnostics") function in the console after loading recommendations to see if there are any problems. Get in touch with Support if there are issues.

๐Ÿ‘

Use LiftIgniter's free Code Review Service before launch

We at LiftIgniter want to make sure that the tracking is launched correctly. After you have reviewed it from your side, please feel free to send it over to Support for a code review so that we can highlight any issues before launch.

๐Ÿ‘

Check our Lab

Once you are ready to deploy to production on a non-negligible slice of traffic, log in to the LiftIgniter ONBOARD page.

Prior to deployment, you should see a pageview count that represents your pageviews in the last 24 hours, but the counts for recommendation sets and clicks should be close to, but a little greater than zero (theyโ€™ll be nonzero because of the requests you made when testing).

Keep in mind that these numbers are only estimates and donโ€™t do some of the backend deduplication we do. Also, since we start collecting widget-specific data only after launch, we may not necessarily outperform the control right off the bat.