Intelligent Decisioning & Open-Source Analytics with SAS CI360 [Hospitality Industry Demo]
In this article and video demonstration, we will cover:
- A customer experience in the context of hospitality, dynamic pricing and offer treatments.
- The demo story will be based on a fictitious hotel called SAS Hotel and a prospective customer is looking to book a stay at the resort property. SAS Customer Intelligence 360 will collect data from the consumer’s website activity, detect a specific interaction event of significance, and trigger a real-time call to SAS Intelligent Decisioning to obtain a recommended treatment driven by a combination of machine learning and business rules.
- After the customer experience demonstration, we will share transparent details on how the personalized treatments are derived and delivered.
Regardless if your brand uses one or multiple solutions to orchestrate communications across touchpoints, there is an opportunity to optimize interactions with target audiences over channels like web, mobile app, email, direct and more. The output of the insights derived from machine learning can specifically inform personalization strategies and treatments. As marketers, we want to provide offers that are both relevant to the customer and beneficial for the company.
Figure 1: Confident decisions at every moment
Machine learning and predictive analytics involve a variety of tasks, such as data cleansing, feature engineering, identifying variable importance, model selection and hyperparameter tuning, which can be tedious to perform manually. Although SAS offers many user benefits in accelerating towards the creation of analytical assets and IP through no-, low- and high code interfaces leveraging the SAS language, analysts who prefer Python or R are not excluded. Analysts can bring in open-source models into SAS, compare them with SAS models, govern and finally deploy through SAS.
Speaking about deployment. Although considered by many as the last phase of the analytical lifecycle, it is also one of the most important. What companies and brands expect from data scientists and marketing analysts is to get all these models in production. The connectivity of SAS Viya and SAS Customer Intelligence 360 showcases an end-to-end approach to monitor and evaluate modeling performance, create workflows to use these models according to specific business rules, and how we can expose these models for both batch and real-time use cases.
A key value proposition we want audiences to take away from this demo & presentation is the ability to use SAS Viya to build modeling projects leveraging SAS, Python, R or TensorFlow. As for the sub-topic of retraining modeling projects as needed, whether the model is from SAS or open-source, the customer experience has no patience for stale model scores and irrelevant treatments. SAS Model Manager ensures SAS Intelligent Decisioning continues to be relevant for personalization & targeting through system-driven champion model retraining or challenger model replacement due to drifting diagnostics of the model’s predictive precision, or feature importance.
The value proposition of any open-source or SAS analytical model assisting in the delivery of targeted treatments needs to ensure:
- The champion model and associated scoring is accessible via open Rest APIs
- Deploy in-batch, streaming, cloud or edge devices
- Support multiple publishing & container destinations
- Monitor, detect and alert users or teams of model or feature degradation
- Retrain the existing champion model as data refreshes over time, or replace the model entirely when a challenger begins performing better
It's easy to recognize that every brand is customer obsessed, and each prospective interaction provides a potential opportunity to make an intelligent decision, deepen engagement and positively impact customer lifetime value. With that said, let's jump into the presentation.
Learn more about how the SAS Platform can be applied for customer analytics, journey personalization and integrated marketing here.
