Acquia CDP
Machine Learning
Machine learning enables software applications to accurately predict various outcomes. It models historical data as input to predict output values. As these models process more samples over time, their performance improves.
Machine learning is classified into two main types:
Supervised learning involves building a model that makes predictions based on labeled datasets, where each input is paired with a corresponding correct output. In the model, both input and output datasets are used to train the system. Based on these datasets, the model can make predictions for similar input data.
The following models in Customer Data Platform (CDP) use supervised learning:
Unsupervised learning extracts patterns and structures from data without labeled responses. In this method, input datasets lack predefined outputs, which allows the model to identify inherent groupings or associations.
Clustering is an unsupervised learning technique that identifies patterns and places similar data points into grouped clusters.
The following models in CDP use unsupervised learning:
The following table lists the machine learning segments:
| Name | Description |
|---|---|
| Likelihood to Pay Full Price - Today | The probability that a customer pays full price for a purchase in the near future. This prediction is derived from analyzing the customer behavior today, such as their transaction history, web activity, and email interactions. For more information, refer to Likelihood To Pay Full Price. |
| Likelihood to Pay Full Price - 1 Month Ago | The probability that a customer pays full price for a purchase in the near future. This prediction is derived by analyzing and learning from the customer behavior a month ago, such as their transaction history, web activity, and email interactions. For more information, refer to Likelihood To Pay Full Price. To surface this model, contact your Customer Value Manager. |
| Likelihood to Pay Full Price - 2 Months Ago | The probability that a customer pays full price for a purchase in the near future. This prediction is derived by analyzing and learning from the customer behavior two months ago, such as their transaction history, web activity, and email interactions. For more information, refer to Likelihood To Pay Full Price. To surface this model, contact your Customer Value Manager. |
| Likelihood to Pay Full Price - This Month |
The probability that a customer pays full price for a purchase in the near future. This prediction is derived by analyzing and learning from the customer behavior in the current month, such as their transaction history, web activity, and email interactions. For more information, refer to Likelihood To Pay Full Price. To surface this model, contact your Customer Value Manager. |
| Likelihood to Buy - Today | The probability that a customer pays full price for a purchase in the near future. This prediction is derived by analyzing and learning from the customer behavior today, such as their transaction history, web activity, and email interactions. For more information, refer to Likelihood To Buy. |
| Likelihood to Convert - Today | The probability that a non-customer makes a purchase in the next 30 days by learning from the customer behavior today, such as their transaction history, web activity, and email interactions. |
| Predicted Lifetime Value - Today - Decile | The predictive lifetime value, segmented into distinct relative decile groups. For more information, refer to Likelihood To Buy. |
| Predicted Lifetime Value - Today - Revenue Group | The predictive lifetime value, segmented into distinct relative decile revenue groups. |
| Likelihood to Engage on Email - Today | The probability that a customer opens an email in the near future by learning from the customer email events today. For more information, refer to Likelihood to Engage. |
| Next Best Channel | The next-best channel to engage a customer on in order to maximize the likelihood of a purchase. For more information, refer to Likelihood to Engage. Some examples are Google and Yahoo. The values are NULL, 0, and 1. |
| Second Next Best Channel | The second next best channel to engage a customer on in order to maximize the likelihood of a purchase. |
| Next Best Channel - Past Score | The past engagement score between 0 and 1 for the predicted next-best channel. |
| Second Next Best Channel - Past Score | The past engagement score between 0 and 1 for the predicted second-next-best channel. |
| Next Best Channel - Predicted Score | The probability of engagement for the predicted next best channel. |
| Second Next Best Channel - Predicted Score | The probability of engagement for the predicted second next best channel. |
| Were recommended a product | The customers who received product recommendations from the CDP’s product recommendations engine. |
| Name | Description |
|---|---|
| Behavior Based Cluster - Today | The customer’s personas based on today’s purchase behavior, preferences, and spending patterns. |
| Product Based Cluster - Today | The customer’s personas based on the products or product categories that customers purchase. |
| Most Likely Cluster | The cluster to which the customer belongs with the highest probability. CDP matches the customers to clusters based on purchase behavior, preferences, and spending patterns. |
| Second Most Likely Cluster | The cluster to which the customer belongs with the second highest probability. CDP matches the customers to clusters based on purchase behavior, preferences, and spending patterns. |
| Third Most Likely Cluster | The cluster to which the customer belongs with the third highest probability. CDP matches the customers to clusters based on purchase behavior, preferences, and spending patterns. |
| Most Likely Cluster Probability | The probability of a customer belonging to a cluster with the best match. Customers are matched to clusters based on purchase behavior, preferences, and spending patterns. |
| Second Most Likely Cluster Probability | The probability of a customer belonging to a cluster with the second best match. Customers are matched to clusters based on purchase behavior, preferences, and spending patterns. |
| Third Most Likely Cluster Probability | The probability of a customer belonging to a cluster with the third best match. Customers are matched to clusters based on purchase behavior, preferences, and spending patterns. |
| Optimal Email Send Time - Overall | The optimal time to send emails for the highest chance of customer engagement, including weekdays and weekends. |
| Optimal Email Send Time - Weekday | The optimal time to send emails for the highest chance of user engagement during weekdays. |
| Optimal Email Send Time - Weekend | The optimal time to send emails for the highest chance of user engagement during weekends. |
If this content did not answer your questions, try searching or contacting our support team for further assistance.
| Likelihood to Pay Full Price - This Month |
The probability that a customer pays full price for a purchase in the near future. This prediction is derived by analyzing and learning from the customer behavior in the current month, such as their transaction history, web activity, and email interactions. For more information, refer to Likelihood To Pay Full Price. To surface this model, contact your Customer Value Manager. |
| Likelihood to Buy - Today | The probability that a customer pays full price for a purchase in the near future. This prediction is derived by analyzing and learning from the customer behavior today, such as their transaction history, web activity, and email interactions. For more information, refer to Likelihood To Buy. |
| Likelihood to Convert - Today | The probability that a non-customer makes a purchase in the next 30 days by learning from the customer behavior today, such as their transaction history, web activity, and email interactions. |
| Predicted Lifetime Value - Today - Decile | The predictive lifetime value, segmented into distinct relative decile groups. For more information, refer to Likelihood To Buy. |
| Predicted Lifetime Value - Today - Revenue Group | The predictive lifetime value, segmented into distinct relative decile revenue groups. |
| Likelihood to Engage on Email - Today | The probability that a customer opens an email in the near future by learning from the customer email events today. For more information, refer to Likelihood to Engage. |
| Next Best Channel | The next-best channel to engage a customer on in order to maximize the likelihood of a purchase. For more information, refer to Likelihood to Engage. Some examples are Google and Yahoo. The values are NULL, 0, and 1. |
| Second Next Best Channel | The second next best channel to engage a customer on in order to maximize the likelihood of a purchase. |
| Next Best Channel - Past Score | The past engagement score between 0 and 1 for the predicted next-best channel. |
| Second Next Best Channel - Past Score | The past engagement score between 0 and 1 for the predicted second-next-best channel. |
| Next Best Channel - Predicted Score | The probability of engagement for the predicted next best channel. |
| Second Next Best Channel - Predicted Score | The probability of engagement for the predicted second next best channel. |
| Were recommended a product | The customers who received product recommendations from the CDP’s product recommendations engine. |
| Name | Description |
|---|---|
| Behavior Based Cluster - Today | The customer’s personas based on today’s purchase behavior, preferences, and spending patterns. |
| Product Based Cluster - Today | The customer’s personas based on the products or product categories that customers purchase. |
| Most Likely Cluster | The cluster to which the customer belongs with the highest probability. CDP matches the customers to clusters based on purchase behavior, preferences, and spending patterns. |
| Second Most Likely Cluster | The cluster to which the customer belongs with the second highest probability. CDP matches the customers to clusters based on purchase behavior, preferences, and spending patterns. |
| Third Most Likely Cluster | The cluster to which the customer belongs with the third highest probability. CDP matches the customers to clusters based on purchase behavior, preferences, and spending patterns. |
| Most Likely Cluster Probability | The probability of a customer belonging to a cluster with the best match. Customers are matched to clusters based on purchase behavior, preferences, and spending patterns. |
| Second Most Likely Cluster Probability | The probability of a customer belonging to a cluster with the second best match. Customers are matched to clusters based on purchase behavior, preferences, and spending patterns. |
| Third Most Likely Cluster Probability | The probability of a customer belonging to a cluster with the third best match. Customers are matched to clusters based on purchase behavior, preferences, and spending patterns. |
| Optimal Email Send Time - Overall | The optimal time to send emails for the highest chance of customer engagement, including weekdays and weekends. |
| Optimal Email Send Time - Weekday | The optimal time to send emails for the highest chance of user engagement during weekdays. |
| Optimal Email Send Time - Weekend | The optimal time to send emails for the highest chance of user engagement during weekends. |
If this content did not answer your questions, try searching or contacting our support team for further assistance.
Acquia CDP