Decision Management for Next Best Action Marketing: How to Bring Together Business Processes, Business Rules and Analytics to Delight Your Customers
- Publication Type:
- Thesis
- Issue Date:
- 2019
Open Access
Copyright Clearance Process
- Recently Added
- In Progress
- Open Access
This item is open access.
This thesis examines the concept of Next Best Action (NBA) Marketing and its uses within the greater context of Customer Relationship Management (CRM). The presented complete NBA Framework consisting of 1. Architecture, 2. Analytics and 3. Project Development and Implementation.
The methodology adopted in this thesis is a combination of both theory and praxis. From an architecture perspective the point is made that 21ˢᵗ century enterprises need to group their resources around their customers. NBA by definition is an orchestration of technology, process and people to deliver a differentiated customer proposition. Therefore, business architecture has to lead the way. To this end, the personalisation of propositions requires a structure or hierarchy. The business issues and groups provide such a structure on which the decision logic can be build. On the analytics side the thesis explores the use of adaptive models powered by Naïve Bayes which provide real-time models in the field of marketing. In this context, posterior probabilities are explored with different types of data grouping techniques such as Platt scaling, Binning and Pool Adjacent Violators (PAV). A model to calculate the final customer score for an online proposition is presented. Unlike conventional marketing, NBA marketing is putting the customer at the centre. NBA is using "propositions," i.e. a proposed course of action, to understand what type of problems the customer is trying to solve. Customer Lifetime Value (CLV) is a crucial baseline metric. The thesis is also exploring various NBA delivery methodologies and applications to other areas of the business such as risk. The last part of the thesis focuses on the implementation of NBA Marketing in the form of project management.
NBA marketing enables the organisation to establish a line of sight between its high-level strategic priorities and its more tactical marketing activities. The decisioning hub is the place where all the decision data and other inputs inform the decision logic. Extending CLV increased the enterprise profitability. Also by using a more extensive model to calculate the customer score for a proposition the more relevant the proposition is and the customer derives a unique value out of interacting with the organisation. This paper shows that Naïve Bayes (NB) is a very appropriate model to work out the customer affinity for a particular marketing outcome. The binning method used in the data analysis is also relevant and appropriate in building the NB classifier. All four experiments performed on a banking dataset show to Binning and PAV are robust methods of grouping the data for analysis. Binning is a robust way to deal with both numeric and non-numeric data. The bins create an ordinal view of the data. Once the data variables are binned Coefficient of Concordance (CoC) is used to measure how good the model discriminates between positive and negative cases. CoC is a convenient way of measuring the power of discrimination of the model because its high granularity and because is cut-off independent. From a project implementation perspective, it is shown that NBA implementations allow the business the change the way they do business rather than an IT driven re-implementation of how the business is currently done. A centralised decisioning system can provide a balanced NBA in real-time across multiple channels will ensure consistency between customer needs and business objectives. Real-time adaptive models should support the decision system by scoring customers affinity for a particular outcome.
Please use this identifier to cite or link to this item: