Global Bank: “Manual Reconciliation”

COMPANY

NAME: Global Banking & Financial Services Company
SIZE: 100,000 employees in 70 countries
INDUSTRY: Financial Services

CHALLENGES:

  • Problems in reconciliation of same transaction being passed over and presented in multiple data streams
  • Analysts were manually identifying duplicate and inconsistent entries in excel preadsheets of security sales data.
  • Data lacked standard naming onventions and other attributes like client names, status, unique ID
Situation: Manual Reconciliation
Help client reconcile data from various different data streams and identify outliers which cannot be identified. Two data streams with similar transaction information but data is presented differently and there is no unique identifier.

Solution: DRaaS + SmartJoin
Multilevel semantic join based on Hadoop i.e. two hoops of jump and then get joins from there. The schema is for ‘MultiJoin’. Configurable rules engine for validation of incoming and outgoing data and create rules for project

Estimated Savings: Approximately 25% resource savings identified as currently 75+ people are doing this exercise manually

Large Bank: “Build Customer Master Data”

COMPANY

NAME: National level securities de-materialization institution
SIZE: Handles ¾ of securities held & settled in dematerialized (demat) form in Indian capital market
INDUSTRY: Financial Services

CHALLENGES:

  • Poor quality of Data
  • Multiple ways of association
  • Large volumes and multiple stakeholders in each account
  • Depository participants to provide know your customer proofs
Situation: Federal Bank requirements require account holders data to be correct, complete, standardized, up to date and de duplicated to ensure accurate reporting.

Solution: Used SmartJoin™ + DRaaS to build customer master data. Included data discovery, cleansing and deduplication along with survivorship, manual validation and updation. Added innovations in tool set to adapt to legacy data, systems and quality of information, created rule framework and best practices for governance.

Result: Bank could identify + 99% of unique customers. Helped bank generate high quality reports, comply with data related regulations. Help embark on journey of using customer master data for cross selling and up selling

Asset Management Company: “Improve data quality, comply with KYC norms”

COMPANY

NAME: Global Asset Management Company
SIZE: –
INDUSTRY: Financial Services

CHALLENGES:

  • Increase in cost of acquiring new customers
  • Crowded & competitive industry with little differentiation in offerings
  • Existing customers facing client manager turnover
  • Focused product marketing with profile of customers needed to identify risk averse and cautious customers
  • Existing CRM affected by poor data quality
  • Consolidated view of client not available. All data managed at single fund subscription level
Situation: Increase existing customer base while maintaining margins. Data fragmented and of poor quality impedes client service & retention management.

Solution: Hexanika was provided limited data (names, addresses, bank details, product type). Provided geographic profile of customer by breaking up addresses into granular elements and profiling customers as urban or rural and further profiling customers based on neighborhoods within urban agglomerates.

Result: Provided a 52 week deduplication service with zero downtime for last 6 years while maintaining, managing and providing customizations for over 70 million records.

Global Bank: “Use Data Quality for KYC management & channel strategy”

COMPANY

NAME: Midsize bank
SIZE: 100 branches
INDUSTRY: Financial Services

CHALLENGES:

  • De-centralized banking
  • Lack of availability of common customer view. Know your Customer (KYC)
  • Data spread across rural, semi urban and urban geographies
  • Day to day business relied on personal rapport. Clean data not a requirement and hence poor data quality
Situation: Client’s aggressive growth transformation and execution of strategic roadmap entailing customer, product and credit re-alignment, process centralization and technology up gradation. Required KYC compliance of all accounts across the Bank to prevent freezing of non compliant accounts; branch rationalization to determine which branches to retain; and consolidated view of customer to migrate data to new core banking system and for CRM.

Solution: Cleansed, standardized, enriched and de duplicated account data; provided geographical boundaries; created unique customer/household views; classified records on address, contact & identity data; identified accounts for KYC information and developed channel strategy based on addresses/contact data at household level

Result: Determined channel strategy for data collection and set the stage for bank’s expansion strategy.

Securities Institution: “Enable compliance; monitor suspicious trends”

COMPANY

NAME: National level securities de-materialization institution
SIZE: Handles ¾ of securities held & settled in dematerialized (demat) form in Indian capital market
INDUSTRY: Financial Services

CHALLENGES:

  • Poor quality of Data
  • Multiple ways of association
  • Large volumes and multiple stakeholders in each account
  • Depository participants to provide know your customer proofs
Situation: Record linkage for meeting compliance. Identify associated parties with multiple demat accounts. Multiple demat accounts were being used for multiple applications resulting in increased chances of allotment of IPO subscriptions

Solution: Hexanika’s expertise, tool sets and services were used to define linkages and association rules for identifying parties within legacy data. Customer data was matched against UN sanctions list; adapters were implemented to convert and normalize incoming data formats to standard formats and reject bad data using business rules.

Result: Customized dashboards were created to enable regulatory compliance.Reports were fed back into depository participants (DP) by client. For specific customers “Know your customer” proofs were provided by DP to complete the cycle

Credit card processing company: “Removal of duplicates in online screening”

COMPANY

NAME: Large multinational financial services company engaged with processing of credit cards for a midsize bank
SIZE: 13 million customers
INDUSTRY: Financial Services

CHALLENGES:

  • Maintain turnaround time of under 10 seconds for every new inquiry
  • More than 24 tokens needed to be matched across several rules involving multiple addresses and multiple phone numbers
  • Poor and incomplete quality of data
Situation: Existing system for credit screening loan applications against existing database of 13 million records missed large number of duplicates. Waste of time and resources expended in manual eyeballing effort. Possible extension of credit to high risk

Solution: Hexanika’s expertise was used to match data and seamlessly integrate with client’s loan transaction system. Multiple iterations were made to ensure negligible skips compared to existing system while ensuring maximum relevant matches made against existing customer database, negative individual database and negative company database. A flexible rule based system was implemented for continuous improvement in match quality and there was flexibility to add new fields for new match rules without IT involvement

Result: Relevant matches increased by 60%

European Bank: “Online de duplication for timely, quality credit screenings”

COMPANY

NAME: Global European Bank
SIZE: Top 10 global bank by market cap
INDUSTRY: Financial Services

CHALLENGES:

  • Data variation challenge
  • Change of key transaction systems in short time frame despite continuous environment change while simultaneously managing data inflows from legacy and new systems.
  • Ensure business continuity. Round the clock services given notwithstanding public holidays as those were high loan days.
  • Provide good match quality and accuracy without compromising on speed to handle large number of loan applications coming every day
Situation: Mitigate risk of bad loans and accompanying possible reputational risk. Client critical requirement was timely and as ‘accurate as possible’ detection of potential fraudulent applications across personal loans, auto loans, mortgages, secured loans, business loans etc. Identify with high level of confidence similarly sounding names & addresses in good turnaround time to enable downstream processes of credit scoring and application approval

Solution: Provided customized ETL to validate, verify and join disparate structures across 7+ types of loan systems in record time; customized middleware to schedule & manage incoming data based in data type and operation; optimal match methodology; and customizable reports of match results for downstream decision making

Result: Client achieved its stated goals in required time to ensure timely & quality credit screening

Insurance Company: “Value based customer segmentation with k-clustering”

COMPANY

NAME: Private Insurance Company
SIZE: 6 million insurance policy database
INDUSTRY: Financial Services

CHALLENGES:

  • Database with profile & investment behaviour variables-existing & derived. Around 100 variables analyzed for completeness, availability and standard values
  • Narrow variables for customers and agents using distribution analysis
  • Optimization of cluster identification for customer groups with similar behaviour patterns
  • Link agent segmentation tree with customer segmentation tree to understand impact of agent behaviour on customer retention
Situation: Identification of key market segments based on profile and potential for cross selling & up selling of new policy types.

Solution: Identified distinct customer and agent segments for targeting new products. Process involved: data preparation, data enrichment, data quality assessment, data transformation, exploratory analytics, building analysis tables, cluster optimization, k-clustering trials, short listing clustering trial results for policy owners and agents, drill down OLAP analysis & recommendations on which segments to target and which product to target

Result: Distinct clusters identified indicating lifecycle behavior of customers along value chain. Based on this, client designed new marketing & distribution strategy

Insurance Company: “Identify unique customer identity”

COMPANY

NAME: Insurance Company
SIZE: 4 million policy holders
INDUSTRY: Financial Services

CHALLENGES:

  • Identify unique policy holders. Telephone numbers and addresses were interchanged for various records of same policy holders
  • Adopt different matching strategies for urban and rural areas
  • Understand date of birth patters and classify records as high/low integrity
  • Assign correct gender
  • Manage old and new telephone numbers and multiple phone numbers in same field
  • Put in place process for end to end implementation
Situation: Identify unique customers and households to increase operational efficiency, cross selling and understand cumulative exposure across a single entity. Identify additional duplicates not detected by in house IT team. Data spread over urban/rural areas; data quality poor and with variations; and large percentages of matches missed using internal

Solution: Defined various linkages and association rules for identifying unique customers and households. Used cleansing and de duplication and supported by expert eyeballing services to identify outliers. Provided client with set of high and low confidence matches for final certification and upload.

Result: Client achieved its stated target of identifying the unique customers and households and achieved a better understanding of total number of policies held by each individual and household

New Bank: “Merge customer database across 8 different financial entities”

COMPANY

NAME: Commercial Bank
SIZE: 900,000 customers, 500 branches, over 1,000 ATM’s. Total assets: US$2.9 B
INDUSTRY: Financial Services

CHALLENGES:

  • Scale: 900,000 + customers to be migrated to new system
  • Heterogeneous Databases Identify and extract relevant data from 8 host systems
  • Tight time frames: Total one month time frame with tight time window with actual migration
  • Incremental data handling: Manage incremental data added during process of migration
  • Manage risk: Ensure accurate de duplication and no loss of information or customer records
Situation: Client converting from finance company into commercial bank. Data from Group’s 8 different legal entities required to be merged into a single cleansed and unique

Solution: Provided expertise of working with data related issues to identify problem sources and fixes. Identified data related issues, anomalies and exceptions in the data and resolved them. Customized rules generated and provided for various types of data- depending on customer type (corporate or individual), customer location and customer address. Incremental de duplication round to match new records added in interim while cleansing process going on with already matched data.

Result: Client achieved its stated goals in required time-merging various databases into single entity.