Improved marketing spend effectiveness for a large entertainment company
Case Studies:Mu Sigma
Published On: 21 August 2015
What We Did: A structured data driven framework to gain a better understanding of the customer base leading to better engagement
The Impact We Made: The improved customer segmentation helped create the basis for a $20 MM additional annual revenue
Summary - Improving ROI of marketing spend
The client’s existing methodology to drive marketing spend on target customers was based on limited data from past behavior. Mu Sigma helped improve the ROI on player reinvestment by helping them gain a better understanding of their customers. This involved the creation of comprehensive customer segmentation and a scoring mechanism to predict a customer’s next trip spend bucket. The new approach resulted in improved accuracy of average daily worth per trip by almost 10% over the base approach in place.
About The Client - An entertainment company
The client is the world’s most diversified casino entertainment provider and an industry leader in establishing a customer loyalty program.
The Challenge - Sub-optimal allocation of marketing spend
The client’s existing methodology to predict customer’s Average Daily Worth (ADW) was primarily based on customers’ prior trip spend. This resulted in sub-optimal allocation of marketing spend, low ROI and wrong set of customers being rewarded. A big proportion of the client’s annual marketing budget of a billion dollars was spent on targeting programs based on their perceived ADW. The client felt that there was room for improvement in extracting ROI out of this marketing expenditure.
The Approach - Customer segmentation and scoring framework
A structured framework was developed comprising of the following:
In contrast to the current approach, a new definition of a ‘customer trip’ was proposed to account for frequent customers visiting multiple properties in the same trip
A customer segmentation scheme was developed based on the average daily potential spend value and volatility
A predictive model was built to calculate the probability of customers in each segment to fall into any of the pre-defined spend buckets
Inaccurately predicted customers from the predictive model were observed to be belonging to the same cohorts represented by certain zip codes. Reclassification algorithm was created using these zip codes
Scoring framework was created for the customers so the learning could be operationalized
The Outcome - Improved accuracy and increased revenue
The project resulted in improvement of accuracy by over 10% through a rigorous and algorithmic approach. This approach is easy to scale across geographies. This process is now being leveraged to get better estimates for all the client’s markets
On an average, a total of 14% of the gaming population is getting a revised re-investment. This represents a $10MM cost savings and an estimated $20MM in additional revenue for the client