Sales and demand forecast for drug manufacturer

Client

A leading India-based drug manufacturer.

Opportunity

The client wanted accurate sales and demand forecasts to reduce costs and streamline their distribution processes. They also wanted their sales professionals to make informed, data-driven decisions to exploit opportunities.

The client's sales team had two main groups, sellers (on the field sales professionals) and sales managers and executives (indirect sales teams). Each group required different levels of sales information and had different KPI’s and insights which they required from the sales forecasting models.

Solution

SG Analytics' team helped the client's sales team obtain powerful analytics-based insights and actionable recommendations by consolidating and applying predictive analytical models onto the fragmented data from individual sellers and the client's CRM database.

The project was split into two main areas:
  • SG Analytics' team started with data gathering, categorizing, standardizing, and cleansing from all the sellers to create a common database of all sales opportunities. For each opportunity, the sellers indicated if it was committed, committed at risk, uncommitted or uncommitted positive. SG Analytics' team also had access to various descriptive data from the sellers and the CRM database, describing all the opportunities and their contacts, product-level information, estimated close date, or estimated amount etc.
  • Next, SG Analytics' team created a sales forecasting model that predicted the probability of wins/losses. The forecasting model was trained with accurate historical data, ongoing feedback, pipeline data, and comparisons of what was forecasted versus what happened. If the model gave wrong predictions, it was used to feed model with the correct data to improve the accuracy and telemetry. The model also applied business logic and best practices to ensure that results are within threshold values and to avoid potential duplicates.

SG Analytics' team utilized Latent Semantic Analysis, Neural network algorithms, and regression analysis to create an additional sales forecasting opportunity tracker which was continuously refined with the feedback from the client's sales teams. The model helped with opportunity management, pipeline management, and forecasting.

Tools used:

Python
SQL
Power BI
Salesforce

Value Delivered

►
1
Increased transparency which resulted in increased productivity and collaboration within the sales team.
►
2
Increased insights on strategic opportunity prioritization, to target the ones which were low hanging fruits first by the sellers.