CDP Scenario 2: Predictive Analytics

This is the second in a series of posts outlining key business use cases for Customer Data Platform (CDP) technology.  As subscribers to RSG's CDP vendor evaluations know, most vendors offer reporting and dashboard services as a backbone for analyzing unified customer profile information. Some then go a step further to provide predictive analytics.  

Reporting and Dashboards

Most CDPs provide native reporting and dashboard services. Advanced reporting capabilities include customizable reports that you can progressively drill-down to gather deeper insights. Some platforms also integrate with external analytics services for more advanced reporting. Most default reporting is really about “what already happened” and not necessarily about “what will likely happen?

Enter Predictive Analytics

Predictive Analytics helps you answer the latter question by predicting some aspect of your overall customer journey.

These predictions can be based on historical data, user behavior, and even live data. The system uses all these to predict a future outcome, such as whether or not a customer is likely to buy, or to recommend “next best actions” within an interactive experience.

Use cases for predictive analytics can come in marketing (e.g., lead scoring via predictive lead scoring models), sales (e.g., recommending specific colored socks to people who bought trousers), procurement (e.g., how much to stock and when to order replenishments), predicting process failures, and more.

CDP tools that offer this service will typically apply machine learning and artificial intelligence-based techniques. Statistical techniques such as logistic regression, data science concepts, and neural networks can get deployed to model your processes and then based on that, the system suggests or recommends action items. With time, and more data, the system is supposed to keep learning and improves its recommendations.

Then Again, Maybe Not

Of course, this also adds complexity to these products. Several CDPs can get you up and running within days but most products that support predictive analytics will need you to do your homework well. You will need to spend time thinking about your predictive models, what is it that you want to be able to predict, and how do you get enough data to test and adapt your models.

Also, you may already possess more advanced predictive analysis services in other tools within your organization that could prove more mature and adept at this.

RSG can help jump-start your analysis. If you are a subscriber, you can use our RealQuadrant Shortlist Generator to find out which CDP vendors excel at this scenario. Our evaluations also call out specific capabilities for this scenario for each vendor.


Other Customer Data Platforms posts

Webinar: Composable CDPs - Myths and Realities

Register to learn what composability really means in a CDP context and how to distinguish vendor fluff from genuine opportunities. Come away with some conclusions that might surprise you!