Executive Summary

AIMIA, headquartered in Quebec, Canada, is a global leader in customer engagement and loyalty solutions for leading retail, CPG, travel & hospitality, financial services and entertainment brands. AIMIA’s solutions help brands identify where their customers fall in key milestones along the customer journey, influencing touch points along that journey. A new methodology called SmartJourney® helps predict gaps in this journey, identify at-risk revenue or new growth opportunities. Building such complex loyalty solutions can be a very daunting effort involving business, data science, engineering and IT teams. It requires building sophisticated machine learning models, iterating them with the right datasets, deploying them in their customer environments and eventually monitoring them in production.

H2O Driverless AI has been able to reduce the model development time for AIMIA in half, at the same time delivering 700% increase in cost savings for their customers’ campaigns. This solution has been a win-win for AIMIA as well as for their customers.

Challenges

Use-cases such as predicting customer churn and fraud detection have been tough code to crack for marketing solutions providers, primarily due to lack of prevalence of the relevant datasets and the steady evolution of fraudulent behavior by bad actors. Developing machine learning techniques in the face of these is challenging. There was a need to increase the agility of model development, build newer use-cases quickly, iterate on them faster, improve overall trust in AI by making the results of machine learning algorithms transparent to business stakeholders, as well as benchmark the performance of the models already in production.

Solution Powered by H2O Driverless AI

AIMIA’s SmartJourney® methodology analyzes customer behavior and engagement across a variety of stages. Using Driverless AI, they are able to develop a classification model that will predict whether a certain customer will churn or not. Automating the complex feature engineering process to derive new features from existing variables significantly contributed overall time-savings. In addition, visualizing the datasets before starting model building, understanding the results in different formats and the confusion matrix made it convenient to derive meaningful insights from the data in a few hours, as opposed to days or months.

Results

1. The predictive churn model allowed AIMIA’s customers to intervene early, re-engage and prevent the churn from happening. This has become an integral capability offered as part of their SmartJourney® methodology.
   
   
2. A global CPG company, an AIMIA customer, was able to deploy this model in a retention campaign and achieve an 11% overall improvement in ROI while the traditional rule-based model was unnecessarily giving out 7X more offers to customers who wouldn’t have churned, the latter also being a huge cost savings for the CPG company.
   
   
3. AIMIA is able to make their predictive marketing capabilities an agile team-wide effort with the following additional benefits:
   1. Ability to turn-around new loyalty solutions much faster than before.
      
      
   2. Benchmark the accuracy of the existing models with the ones developed using H2O Driverless AI.
      
      
   3. Communicate the results with the rest of the business and operational teams using the unique Machine Learning Interpretability (MLI) capability in H2O Driverless AI, establishing much-needed trust in AI as a result.

Next frontier in AI for AIMIA

With the predictive churn model now tested and proven out with their clients, AIMIA is confident that they can use the Driverless AI platform for similar advanced use-cases critical to their business. They are looking to build new models in fraud detection and identify new growth opportunities for their customers. This will help mitigate risk and contribute to the topline for the company over time.