Datanyze Customer Success

Take your Datanyze skills to the next level. Watch videos, learn best practices, enjoy training materials, connect with others, share ideas and request support.

Predict Best Practices

Prerequisites

You must have Predict enabled.

Predict helps you identify new leads that share common positive and negative attributes with your existing list of customers or perfect prospects. You can use Predict to uncover traits your existing accounts share and identify new accounts that look just like your best customers.

What's in this article?

  • Building your first model
  • Build targeted predict models
  • Using predict in targeting
  • Syncing predict grades & scores to CRM or MA tool
  • Customize predict grade & score thresholds

Your First Predict Model

1. Click on the Predict module in your Datanyze account and select Create new model.
2. Choose whether to upload a CSV of domains or Use Salesforce Data
     a. CSV - upload only a list of domains
     b. Salesforce Data - requires you to connect via oAuth 1.pngPro-Tip: For first time Predict users, we recommend uploading a list of at least 100 companies. These should be your perfect prospects or perfect customers.

3. To create a model:
     a. Give your model a name
     b. Upload your accounts
     c. Upload your prospects (optional) 2.png

4. Click Generate model.
5. You will be notified that your model is being created. You can monitor the status via the Status bar. 3.png6. You will receive an email when the model is complete and the status will change to Complete in the platform

4.png

7. Once the model is complete, take a close look at the model to view the results. You can view the scores and grades assigned to your Existing leads uploaded, and click on the New leads tab to view net-new leads generated from the model. 5.png

8. Click on the Summary tab to view the Positive and Negative attributes.

6.png

9. This is where machine learning comes in. Thumbs up or Thumbs down the results. Doing this allows teaches your model by telling us which leads are great and which are not. The model will be more refine the next time you run it.

7.pngTargeted Predict Models

1. After you have mastered your first Predict model, you can get more creative by building models based off company size, vertical or other characteristics.
2. Segment your current customers into SMB, Mid-Market and Enterprise accounts and build a model for each market.

8.png 

3. Sync over the model grades and scores into Salesforce so reps can easily access them.

Using Predict in Targeting

1. In the Targeting module, choose the condition Prediction Model, select your model and choose which grade(s) you would like to include in your report.

9.png

2. Further filter the results by applying additional conditions, or filter out Salesforce/Hubspot data using the Tag condition.

10.png

3. Export the results to CSV or search for Contacts.

11.png

Syncing Predict Grades and Scores to CRM or MA Tool

1. You can sync over as many Predict model grades and scores as you would like. Let your CSM or Integrations Specialist know which models you would like to sync.
2. Create fields in your CRM or Marketing Automation tool in the following format for each model you would like to sync:
     a. Datanyze_Customer_Model_Grade__c (Text)
     b. Datanyze_Customer_Model_Score__c (Text)

Customize Your Predict Model Grade and Score Thresholds

1. The default thresholds for predict models are seen below:

 12.png

2. To update these, go to your Predict model summary and update each threshold.

13.png

3. Updating the thresholds can help reps prospect into accounts they would otherwise view as not qualified, if the lead showed an A vs. B grade lead, for example.

Was this article helpful?
1 out of 1 found this helpful
Have more questions? Submit a request