Driving Growth: Enhancing Sales and Customer Engagement with ML

Business Opportunity

Improving Sales Forecasting Accuracy

The client is an automobile OEM manufacturer in India. The brand has been well-received in India. They are utilizing SAP for their dealer management (DMS), ERP, and CRM applications, with data distributed across both SAP Cloud and on-premises solutions. They were looking to streamline their data ingestion from SAP into a centralized cloud-based data lake and improve sales forecasting accuracy, thereby enhancing operational efficiency and decision-making for around 90 dealers.

Niveus created a data lake solution to improve the efficiency of SAP Analytics Cloud, which was previously used to generate and email Excel reports to over 90 dealers. By separating storage from analytics, they achieved better scalability. Additionally, Niveus developed a dealer sales forecasting ML model and a PowerBI dashboard to boost data analytics capabilities. Our solution aimed to predict the probability of a customer purchasing an automobile from the client using a machine-learning classification model. By analyzing historical data on customer interactions, we intend to provide insights that can enhance sales strategies and customer engagement.

Technical Challenge

Ensuring data quality and identifying the right predictive model was the key technical challenge here. The data for this project was sourced from Google BigQuery, utilizing three primary tables:

Our Solution

Niveus built a data lake solution to address the inefficiencies of using SAP Analytics Cloud to generate and email Excel reports to over 90 dealers. They separated storage and analytics for better scalability. Additionally, Niveus developed a dealer sales forecasting ML model and a PowerBI dashboard to enhance data analytics capabilities.

We combined data from the Opportunity, Leads, and Appointments tables to form a comprehensive master dataset. To prepare the data, we consolidated and encoded categorical variables and handled missing values and outliers. We analyzed feature distributions, visualized relationships between features and the target variable (customer purchase), and created correlation plots to identify multicollinearity.

Model Training

Train-Test Split: The dataset was divided into training and testing sets.

Initial Model Trials: Tested Logistic Regression and Decision Tree algorithms.

Final Model Selection: We chose the Random Forest algorithm for its superior performance, fine-tuning its hyperparameters for optimal accuracy and generalization.

Results

The Random Forest model, with an accuracy of 98%, provides valuable insights for the client to enhance their sales strategies and customer engagement efforts. By predicting the probability of customer bookings, the client can tailor personalized advertisements, invitations, and phone calls to potential customers, thereby increasing the likelihood of conversion. Insights from the model, such as the importance of test drives, follow-up frequency, and customer demographics, can help optimize sales strategies to focus on high-conversion opportunities. 

Understanding factors influencing customer decisions allows for targeted engagement strategies that improve overall customer satisfaction and loyalty. Based on the findings, the client could further explore integrating real-time data and feedback loops to continuously improve model accuracy and adapt strategies in response to changing market dynamics.

Benefits of Building ML Models with Niveus

• Enhanced Sales Strategies: By predicting the probability of customer bookings, the client can tailor personalized advertisements and marketing campaigns, focusing on customers most likely to convert. Sales teams can prioritize leads with the highest potential, allowing the client to optimize their efforts and resources on high-conversion opportunities.
  
• Data-Driven Decision Making: Understanding the factors that influence customer decisions, allows the client to make informed decisions based on data rather than intuition, leading to more effective strategies. Integrating real-time data and feedback loops enables the client to continuously improve model accuracy and adapt strategies to changing market dynamics.
• Increased Conversion Rates: By concentrating efforts on leads with the highest likelihood of conversion, the client can significantly improve their overall conversion rates. Identifying key drivers of customer behavior can help the client refine their sales process, making it more efficient and effective.

Power of Partnership

Partnering with a Premier Google Cloud Platform (GCP) partner like Niveus Solutions brings significant advantages. By leveraging their deep expertise in GCP services and industry experience, businesses can implement robust, scalable, and secure solutions. Niveus Solutions offers access to advanced tools and custom solutions, streamlining cloud adoption and optimizing resources for enhanced operational efficiency. Their continuous support and strategic guidance ensure long-term success, while their focus on security and compliance builds trust. This collaboration fosters innovation, agility, and flexibility, enabling businesses to quickly adapt to market changes and maintain a competitive edge.