You can view the status of a predictive model from the Predictive studio index page.
If data requirements are not met, predictive scores will fail and there will be an alert on the studio dashboard and an error on the index screen. If the model fails, consider tuning the prediction window, goal lookback days, user attribute filters.
The status of a predictive model is indicated by various colors. Based on the status, you can take actions such as tweaking the definition or resolving errors.
Here’s how a model status is color coded and what it means:
| Green | Scores are being computed as expected and you can use the scores based on your requirements. |
| Red |
An error occurred while computing the scores. An information icon is displayed. Hover over the icon to see the type of error. |
| Light grey | Scores are not ready or computing is paused.
|
| Dark grey |
Model is archived. Computation is stopped for archived scores, but you can use the filters to see them. You can activate an archived score by using the drop-down menu. For more information, see Archived entities. |
Open the predictive model to find more information about the error on the model summary tab of the details page.
Here's a list of errors that you can see and possible steps to resolve them.
Data errors
Predictive scores are not refreshed if the new data that the Blueshift platform receives does not meet the criteria specified when you defined the predictive model. Here are the possible cases when it happens.
Insufficient data: This error can occur if the received data is insufficient for the Blueshift platform to compute a score. For example, you have created an SMS engagement model but the customers you choose in the segment don't receive enough SMSes. In such a case, Blueshift cannot compute an SMS engagement score. To resolve this error, select a lookback period in which your customers receive an adequate amount of SMSes.
Insufficient conversion rate: This error can occur if the conversion rate of a campaign is too low. For example, if you are using an email channel to send messages for your campaigns, and in the email messages you send a list of products to your customers. However, if the conversion is less than 1% of the total number of customers to whom you send a message, Blueshift cannot generate an accurate score. This happens because there aren't enough conversions to train the models. To resolve this error, add filters in your campaign so that it targets customers who have purchased your products previously. This will allow Blueshift to use a customer set with more than 1% conversions.
Low predictive power: This error can occur if the data received is noisy. For example, if you select a segment that targets customers in the US but Blueshift receives a lot of data for your customers who reside in Canada. In such a case, Blueshift cannot properly compute a score and pauses the score refresh. To resolve this error, ensure that the data that Blueshift receives is clean and consumable. One of the issues could be improper integration, so we recommend that you review our developer documentation if data from your service is sent to us using our APIs. Or, you can review documentation on S3 imports and exports if you send your data using S3.
Internal error
This happens when Blueshift’s internal systems run into an issue. To resolve this, contact your CSM or contact us on support@blueshift.com.
Comments
0 comments