At PEL, we believe in making strategic decisions backed by data-driven insights and aim to embed the same in day-to-day processes. Our global presence, diversity of businesses and multiple touchpoints give the Company access to a significant amount of data. As we become more proficient at collecting and managing data, the opportunity to find valuabale insights through analytics is ever-expanding. The Analytics function enables the Company to leverage this data and attain competitive advantage across business lines.

A brief summary of tools utilised by the Company across its Financial Services and Pharma businesses is as follows:

Financial Services

Fraud Analytics Rule Engine (FARE) for Retail Housing

The Fraud Analytics Rule Engine (FARE) aims to create a data-driven fraud detection platform. Some of the approaches followed include:

  • Machine-enabled checks for aspects that are currently performed manually, e.g., PAN checks
  • Integration of information from third-party fraud detection service providers, which match loan application data against fraud databases and dedicated ‘watch-lists’

The system ensures minimal fraudulent applications and improves the overall quality of lending. Further, it categorises the applications under various risk heads ranging from ‘Deep Red’ (high risk) to ‘Green’ (least risk). Going forward, we plan to further enhance fraud analytics through collarboration with fintech partners.

Credit Analytics Rule Engine (CARE) in Retail Housing

The Company has developed a proprietary predictive credit risk model for retail housing. This helps assess the probability of an applicant to go over 90 dpd (days past due) in the next 24 months. Going forward, we plan to further enhance the model by incorporating additional credit assessments criteria, such as loan-to-value, fixed obligationto- income ratio, etc.

Account Monitoring Framework for Retail Housing

The Company has developed an account monitoring framework for retail housing. This ranks customers in the order of possible default within a 3-month horizon, by leveraging internal performance, bureau scrub and underwriting score.

Sales and Marketing for Retail Housing

Sales And Marketing Analytics Rule Engine (SMARE) helps in stiching the data across multiple systems to track the end-to-end journey of a prospective customer, starting from lead generation campaigns to the final loan disbursal. This has resulted in improved customer experience and led to higher conversion improvement in client acquisition / conversion rates.

Financial Model for Mid-Market Construction Financing

For evaluating mid-market construction fianancing deals, Analytics helped in the automation of financial modelling to facilitate decision-making for the Deal Clearance Committee. This led to significant time savings for analysts, as well as reduction of manual errors, thereby improving employee productivity.


Global Pharma Products business

For the Global Pharma Products business, a tool has been created to automate the sales reporting process for certain products. The automation ensures a standardised and error-free process and considerably reduces the report creation time.

Another core initiative includes building a model to predict likelihood of purchases using relevant metrics, such as monetary value, repeatability and recency.

Hence, using advanced technologies, the Analytics function is supporting businesses in maintaining and enhancing their competitive advantage.

India Consumer Products

Use of analytics has been instrumental in identification of the following goals for the Consumer Products Division:

  • Categorisation of distributors based on key parameters such as business relevance, risk etc.
  • Smart Selling model, which enables terrirtory sales officers (TSOs) in recommending brands to retailers and driving cross-sales
  • Analytics are being used for enhancing sales through right hiring and retention of well-performing TSOs
  • Optimising spends on Search Engine Marketing (SEM) across multiple e-commerce channels such as Amazon and Flipkart

Machine learning to reduce risk exposure
In the India Consumer Products business, we categorised distributors based on business relevance and risk, to minimise any high-risk exposure. We leveraged a machine learning algorithm called ‘K-Means’ to create heterogeneous clusters of distributors, by utilising in-house data on their performance. Subsequently, results from the exercise were collated to arrive at individualised creditlimits for distributors. Additionally, the movement of distributors across categories is monitored on a monthly basis to reset the individualised credit limits, if required.