Troubleshoot customer insight

  • Updated

This article explains why you may not be seeing customer and product insights about your brand's performance.

Predicted Order Likelihood, Churn Prevention, and Observations/Insights attributes require that a customer has existed for at least a day, as the data updates overnight. If you are missing information for a specific customer, please ensure they have been in the account for at least 24 hours.

Predictive Order Likelihood

In CustomersInsights:


Here are a few reasons you may not be seeing the Predictive Order Likelihood explorer:

Orders are not being sent, or there are too few of them. The model can not make predictions without an adequate number of order events. This situation might happen when:

  • The account is new and does not have enough past data synced.
  • Offline purchases are not sent to the account.
  • The business deals with very low volume.

Work with your Customer Success Manager or our Support Team to sync all possible orders and events to your account.

The model does not have enough consistent data. If all shopping and purchasing behavior is not or cannot be sent consistently, the model cannot understand what happens before a purchase. This situation might happen when:

  • There is a combination of online and offline purchases where offline data is not synced.
  • Offline shopping makes up a significant part of the consideration cycle.
  • The event that indicates a conversion is not an "Order" event.

Work with your Customer Success Manager or our Support Team to see if there are other events you can sync or different types of events to identify as "Order" events.

Product Opportunities


Here are a few reasons you may not be seeing the Product Opportunities explorer:

  • Product detail pageviews are not firing.
  • The product catalog is not synced.
  • Orders are not synced.

Check that your integration is sending the appropriate data. You can work with your Customer Success Manager or our Support Team.

The Product Opportunities explorer will not be available for inventory-1 SKUs (auctions, one-of-a-kind, or custom products).

Churn Prevention/Winback


If you are not seeing the Winback explorer or Winback attributes populated, the model does not have enough data or variation of data. Consider the following:

  • Have you integrated multiple data sources? Winback works best with a minimum of order data and email open data. The model gets stronger as you add more data sources and events.
  • Is there more data you could bring for existing data sources? For example, last open or only opens in the last 30 days will limit the model's ability to identify the engagement cadences for your customers. Longer data makes a better model. 

If you need assistance integrating more data with Optimizely Data Platform (ODP), please reach out to our Support Team.

Observations and insights attributes

Most observations and insights are based on order data, product data, campaign data, or site visitation data. If one of these is missing, the attribute will not populate. If you think you are sending that data and the missing attribute data is in error, or you want to sync additional data, work with your Customer Success Manager or our Support Team to get this data flowing.