Gmid associates services portfolio bank
0 likes329 views
Gmid Associates provides analytics services including predictive modeling, descriptive analytics, data mining, and dashboard solutions. They have experience across industries including banking, insurance, and retail. Case studies highlighted include developing churn prediction models for a telecom company, sales forecasting for an apparel retailer, and implementing collection scorecards for a bank. Gmid aims to help clients make better data-driven decisions through analytics.
Gmid associates services portfolio bank
- 2. Clarity on constantly evolving Business Dynamics Data backed decisions are more contemplative and thus wiser Data Disabled Decisions Data Powered Decisions The Power of Analytics… 2Copyright © Gmid Associates.
- 3. Agenda About the Company Capabilities and Services Predictive Analytics Solutions Other Services Descriptive Analytics Solutions Data Mining & Cleansing MIS / Executive Dashboard/ Simulation Tools Relevant Case Studies 3Copyright © Gmid Associates.
- 4. About the Company 4Copyright © Gmid Associates. Experienced Team – Professionals with decades of International Experience Delivery Across the Globe – Analytics Partner Best Talent – Graduates from IITs, IIMs, ISI Industry Knowhow – Complete Lifecycle of Industry
- 5. Gmid Associates has a global footprint 5Copyright © Gmid Associates. HQ Gmid Rep office Tie ups/networks California London Australia Delhi Mumbai Bangalore New Jersey
- 6. Services 6Copyright © Gmid Associates.
- 7. Analytics is at the core of banking 7Copyright © Gmid Associates. Identification Validation Authentication
- 8. Predictive Analytics Solutions 8Copyright © Gmid Associates. Using statistical techniques –Regression, Time Series, Neural Models etc. on historical information, one can accurately predict the outcomes of future events and use this information to plan preemptively Model Development Framework % Population 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 20% 40% 60% 80% 100 % of Delinquent 80% lift! Model Validation Chart Maximize the divergence between the distributions of target / non target (good / bad) accounts Estimate the unknown population characteristics based on sample information Use propensity scores to calculate probabilities of default, expected loss, intentional fraud etc. Maximize revenues, rationalize expenses.
- 9. Descriptive Analytics Solutions 9Copyright © Gmid Associates. Use statistical clustering schemes, econometric techniques and overlay with business inputs to group ‘identical accounts’ from a pool of customer population and create targeted segments for focused treatment Use segmentation solutions to optimally allocate marketing budgets, increase customer service levels and loyalty, manage bad debts and maximize collections
- 10. •We help organizations transform and combine disparate data, remove inaccuracies, standardize on common values, parse values and cleanse dirty data to create consistent, reliable information Data Cleansing and Enrichment •Your customer contact numbers are buried in a dataset that has all sorts of text entries. This makes contactability on those datasets very difficult •We have tools that dig valid phone numbers from deep into the text data, intelligent enough to complete incomplete numbers (i.e. adding STD codes) •Tools can be customized to suit your business requirements Contactability Improvement Tools •We have proven tools and expertise to run customer de-duplication algorithms on data, identifying unique customers/ households/ relationships and establishing mappings amongst them •This helps you understand your customer data, draw critical conclusions and make meaningful business decisions Data De-duplication Solutions Data Mining & Data Cleansing 10Copyright © Gmid Associates.
- 11. MIS/ Dashboards/ Simulation Tools 11Copyright © Gmid Associates. Measure efficiencies/inefficiencies Ability to identify and correct negative trends Ability to generate new business opportunities Align strategies and organizational goals Save time over running multiple reports Gain total visibility of all systems instantly Enables better decisions making by building monitoring systems that are : • Real time, • Correct, and • Efficient We use advanced Analytical techniques to make sure that the techniques which best captures the business problem is used. The outcome algorithm is built into scenario analyzer tools. Ability to make more informed decisions based on collected business intelligence
- 12. Case Studies 12Copyright © Gmid Associates.
- 13. Outline BenefitsProject Description Monthly default prediction models for active Auto Loan portfolio After implementing the model, the monthly default rates are down by 16% A 100 year old leading vehicle finance company from US with a sub prime portfolio wanted to make scientific and optimal collection strategies. We developed an early warning delinquency predictor scorecard on the portfolio and implemented the same on the client’s system. The scorecard runs on the last day of every month and gives scores to all the accounts based on their propensity to default on the payments in the next month. It also categories the accounts into risk segments, using which the company can make effective, targeted collection strategies ‘Bad’ Application Prediction system for US Auto Finance company Early Write off losses have gone down by 12% within 4 months of model implementation A leading auto finance company from Texas wanted to devise effective strategies to separate good applications from bad ones to improve the portfolio quality and minimize future credit losses We developed two predictive scorecards- write off prediction and early pay off prediction. Since both of these were loss making scenarios for the business, we clubbed them together and devised “high/ Medium/Low” risk bands. Every application is given scores and risk segment and business makes effective acquisition strategies Case Studies: Predictive Analytics 13 Cross-sell Strategies on a diversified financial portfolio Cross-sell penetration increased from ~1% to ~4% on a base of 2,50,000 accounts in 8 months time after implementation A global financial services firm needed to develop effective and easily implementable cross-sell strategies on a diversified portfolio containing mutual funds, life insurance, general insurance and mortgage loans Developed multiple statistical models and used clustering techniques, overlaid with business rules to design efficient cross-sell strategies Copyright © Gmid Associates.
- 14. Outline BenefitsProject Description Collection Scorecards and Strategy for Retail Bank Write-off losses down by 4% within six months. Impact in the range of $5million South India’s leading bank wanted to put in place a centralized collections system powered by predictive analytics. Two statistical models were developed- first on the current portfolio to predict potential payment default cases and second, a payment prediction model on the 90+ DPD portfolio. Segmentation and collection strategy put in place to run preventive campaigns in order to minimize delinquencies at lowest expenses Case Studies: Predictive Analytics 14Copyright © Gmid Associates. Churn Prediction Scorecards for US Telecom Major 85% potential churns captured in top 20% ‘risky’ population 15 days in advance Churn rates down by 7% One of the world’s largest Telcos wanted to replace their existing churn prediction system with a better quality product to arrest increasing churn in their consumer and business portfolio Two fold problem description- early identification of subscribers most likely to churn, identification of important factors driving the churn. Developed predictive models to aid in churn prevention campaigns. Ran statistical tests to identify most important variables that affect churn behaviour. Sales Forecast Model for world’s biggest fancy dress e-retailer 80% accurate forecast figures within 3 months of model deployment Increased service level by about 41% The client, a large textile manufacturer in UK, wanted to build a reliable model that complements or takes over their in house traditional approach to forecasting primarily driven by gut instinct. Developed a multi variable time series forecasting model for all of the product categories. The model uncovered patterns caused by seasonal variations, lead times, minimum quantities, re-order quantities, extraordinary usage (special orders) etc.
- 15. Outline BenefitsProject Description De-duplication of Records on a Multi Product Financial Portfolio The client was able to know exactly how many customers it had As a result of customer-to- account mapping, they are able to make better informed business plans A leading private sector Insurer was having troubles managing their data warehouse (DW) because there was no unique identifier at client level. Every new acquisition was a new relationship at DW level. Developed a de-dup algorithm by using matching algorithms like levenshtein distance, soundex, metaphone etc. Created a new client level UID. Intensity Optimization for a Cards Collections Campaign • The collections unit was able to reduce their costs by 15% for the same collection revenues. A multinational bank’s cards collection business needed help to identify optimum reach out intensity/ channel at bucket and segment levels. Using statistical and econometrics techniques on historical data, we devised optimum channels and intensities at bucket and segment levels. Data Cleansing on a Life Insurance portfolio database • As a result of improved contactability, cross-sells went up by 11% in a year’s time Our client, a leading Insurer needed to increase phone number contactability on their database. The phone numbers were not in machine dialable format s hence dialer machines could not be used directly for customer reach out/ cross-sell/ renewal campaigns. Developed an algorithm which picks correct phone numbers from text strings and modifies them into machine dialable format (adds std code on landlines, if required). Automated it in an excel tool to be used on incremental data. Case Studies: Data Mining & Cleansing 15Copyright © Gmid Associates.
- 16. Thank You Mudit Chandra Sales and Marketing Head Gmid Associates Mobile Number +91-98107-96148 Email [email protected] Web www.thegmid.com Copyright © Gmid Associates. 16