Share this post:

Generally speaking, sales drives everything else in the business – so, it’s a no-brainer that the ability to accurately predict sales is very important for any business. It helps companies better predict and plan for demand throughout the year and enables executives to make wiser business decisions. Not only that, with accurate sales forecasting, businesses can devise cost-effective marketing plans that rally demand during the right times. Another important positive aspect is, with accuracy, it fosters trust amongst employees and shareholders.

An insurance company (referred to as the Client/the Customer henceforth) tasked IBM with improving predictability in sales in 2021 using IBM’s Data and AI platform – Cloud Pak for Data.

Business Challenge:

Market campaigns by the client were not effective as they were targeting their customers based only on the basic demographics and not factoring in the other relevant attributes such as education, habits etc. Also, BDM (Business Development Manager) assignments to the generated leads were not optimally leading to accurate forecasting. Here is a concern cited by their CEO in her own words: “If I have a prospect in Ahmedabad who is in his mid-thirty, a non-smoker and running his own business, I don’t know which BDM (Business Development Manager) will be ideally suited for this lead such that I’ll have higher chances of conversion”.

All they needed was to make the sales forecast accurately based on the various relevant and comprehensive sets of parameters in a given month, grounds from branch level to national level in an online real-time mode.

The Journey:

An MVP (Minimum Viable Product) scoped around the following use-case was undertaken by IBM.

Analysing the sales lifecycle:

• To help predict sales shortfalls
• Understand and quantify factors that affect the execution of an activity by a given BDM

The following CP4D capabilities were chosen to meet the above requirement.

• WML – Watson Machine Learning
• WKC – Watson Knowledge Catalog
• WS – Watson Studio

The following solution flow was finalized by the IBM team.

• The CP4D (v3.5.6) components were installed on OpenShift by the IBM team.

• Once installed, the IBM data-science team built the data sources & data catalogues. After that, they built, deployed, and ran the models to analyse the data as per the flow mentioned in the flow diagram.
• Daily cadence meetings with the customer Analytic team had been set up to co-create the MVP (Minimum Viable Product) in a partially overlapped fashion. This served as a great opportunity for the client team to get first-hand exposure to IBM’s products.
• Later, visualizations were also built using the Embedded Cognos Dashboards to depict the results for easier consumption.

Why IBM:

• Single unified hybrid platform: One unified experience across all data services while providing ease of provisioning and monitoring. It enables data engineers, data stewards, data scientists, and business analysts to collaborate using an integrated multi-cloud platform.
• Speed in model deployment: Watson Studio provides various tools such as model builder, flow editor, experiment builder, notebooks etc. for designing, training, and managing machine learning models faster and at scale.
• Modern Data Governance: Data automatically integrated with data governance capabilities for self-service data discovery
• A truly hybrid-cloud world: Ease of portability across private and public clouds. Run it on-premises, on an IaaS vendor, on managed OpenShift or as a service.

The customer will derive the following benefits from IBM’s Data Platform and its proven Garage Methodologies:

The customer can gain valuable feedback from the visualization of shortfalls in conversion across organization as well as for individual and pivot as necessary. That would lead to:

• Improved sales planning: Assigning pipeline building activity optimally to the BDM (Business Development Manager) thereby optimizing sales cycle according to the strength of BDM.
• Accurate sales forecasting

The Future:

• Following and contingent upon the success of the MVP, the customer is looking to expand by onboarding a team of 15+ data scientists to analyse the data to gain various insights. This will require the expansion of the cluster infra.
• IBM team will also upgrade the current version of CP4D which is 3.5.6 to 4.0.2. The newer version will provide enhancements to SPSS Modeler and other CP4D services.
• As recommended by the IBM team, the customer is now mulling over having a data-lake and warehouse strategy to capture more data and refine it in order to produce the accurate outcome.