Log in

Testing and running recommendation Premium plan

Before running and while the recommendation is active, you can test the event sending and display of offers on test users.

You can open the test page by clicking the button from the recommendations list.

Events vaildation

The Events validation tab displays data for the last five hours — the percentage of events successfully transferred to MyTracker.

Make sure that the following conditions are met for the recommendation to start and function correctly:

  • An offer_trigger event has been passed for each application, that is, the percentage of events passed is greater than 0.
  • For the purchase_offer, offer_card, and account_status the percentage of successfully transmitted events is at least 95%. Smaller numbers will be highlighted in red. If there are errors in the format of the transmitted data, they will be described in an information message.

For more information about events and data format, see the Integrate MyTracker SDK section.

Verification on test users

You can verify the recommendation on test users through your app interface or through requests to the Personalize API.

  1. On the Test users tab, click Add user.
  2. Fill out the form:
    • Application — select the application you will be testing.
    • Offer group — select the control group or one of the recommendation groups to which the test user will be assigned.
    • custom_user_id — enter the user ID. This can be a fictitious ID if you are testing through the Personalize API or a real user ID if you are testing through your application interface. For more information about user identification, see the Integrate MyTracker SDK section.
  3. Click Add user.
  4. Repeat steps 1-3 to add test users for each offer group.
  5. Test the recommendation display through the API or your application interface:
    • Using the API, you can send a recommendation request for a user with the custom_user_id specified for the test. See the Integrate Personalize API section.
    • To test through the interface, you need to log in to the application as the user with the custom_user_id specified for the test.

The test will succeed if each user sees an offer from the group selected for them. You can find the offer list on the recommendation view page.

We recommend that you also test displaying an offer for a user outside the test group (outside the selected segment), for whom the default suggestion should be displayed.

Running a recommendation

  1. In the recommendations list, open the test page by clicking the button .
  2. Check the event sending.
  3. Check the recommendation on test users.
  4. Click Run recommendation.

Users of your app will start receiving personalized suggestions. The recommendation status will change from Testing to Learning. For more information about learning time and next steps, see Monitoring the recommendation.

Was this article helpful?