Data helps a business make better decisions. There are a lot of approaches a business can take to leveraging data. In this blog post, we will walk through a few different approaches to using data which span both Business Intelligence (BI) and Machine Learning (ML).
Analytics, in the context of business decisioning, involves summarizing past data in order to provide information on how to make optimal decisions in the future. For instance, if a business would like to reduce churn on a B2C site, analytics might convey the average page views for both users who churned and users who did not churn. This analytics information could then be used to set thresholds around what page view activity level might lead to user churn.
Predictive modeling takes a different approach. Predictive models typically leverage past data in order to make predictions about the future or unknown aspects of the present. Let’s take again the example of user churn. A predictive model would look at all past behaviors across all users and identify which patterns of behavior were most indicative of user churn. The more sophisticated the predictive model, (predictive models range from logistic regressions to deep neural networks) the more sophisticated the patterns of activity the model can leverage to make its predictions.
Prescriptive models (like Uplift Modeling) focus not on predicting the future, but rather the best actions to impact the future. Prescriptive models “prescribe” the best intervention in order to achieve the desired business outcome. Let’s think about user churn again. A prescriptive model might tell a business what offer to prescribe for each user in order to minimize user churn. A predictive model, on the other hand, would only be able to convey whether or not a user will churn. At Vidora, we often note that many marketing and product-related problems can be thought of in terms of prescriptive modeling.
Cortex offers numerous flavors of Predictive and Prescriptive modeling. Vidora’s Real-Time Decisioning SDK also offers both predictive and prescriptive approaches and can be leveraged to enable real-time decisioning across product experiences.
As a business you have a lot of choices to make around how to leverage data. We are here to help! If you are interested in talking through any of the areas above, please reach out to firstname.lastname@example.org