Lead Scoring

Lead scoring is a method to evaluate and prioritize leads based on the their potential value to the organization.

Predict Lead Scoring

Lead Scoring by Likelihood

This example uses a user table to lead score potential users. The table contains basic information like the origin of the interaction (search, social, direct), the page they landed on, device used, and country of origin. If a user signs up, they will be assigned a user_id and otherwise this will remain NULL.

  1. We SELECT data from the user table to create a CTE called "conversion". In this CTE, we create a new column converted which describes when a user has converted (signed up) or not.
  2. The PREDICT function is used to predict a value for the converted column for each row in the conversion table.
  3. To look at the lead potential, we select the rows where the model predicts conversion (prediction='Converted') but the user has not yet signed up (user_id IS NULL). 
  4. The resulting rows are then ordered by the probability column in descending order. We return all of the columns from the conversion table (* indicates to return all columns), as well as the prediction and probability columns.

The final result shows the users most likely to be converted! Try reaching out!

Understand Lead Scoring

Understanding Lead Scores

This example shows how you can easily understand the drivers of lead scoring.  

It's super simple - all you need to do is wrap your PREDICT function with EXPLAIN ! Infer then uses Explainable AI to examine your lead scoring model, and figure out what the drivers of conversion are.

The final result shows the importance of each column (features) in your conversion model. Now you can understand where is driving sign up and you can use this information to action on future targeting efforts!

Predict Lead Value

Forecasting Lead Value

This example shows how you can forecast the value of a potential lead. 

Here we simply predict the average user spend, for those that have not yet signed up. You could combine this with the previous example by multiplying the probability of signing up with the expected spend to get an overall lead score!

The final result orders the results by largest expected spend, so now you can focus on the customers with the biggest pockets!