Ian Kirton TSB Internal Audit Data Analytics the

Ian Kirton TSB Internal Audit Data Analytics – the TSB journey April 2017 Public

Contents Overview of TSB, our Team and our approach Our data analytics experience so far Future vision and plans 2

TSB - The new bank with a 200 year history 6. 6 million customers • 22. 3 billion in current accounts and savings • 18. 3 billion in mortgages • 1. 3 billion of Loans 8500 ‘Partners’ • 631 Branches • 10 Head office locations • 29 million items of new marketing literature to support launch 3 yr project • May 2011 management team appointed and 5500 branch colleagues moved to create the Verde team • May 2012 recruitment started for core functions • 4400 colleagues involved in project build 3

Establishing the Audit Function 1 Leverage organisational strategy 2 Develop Audit Strategy 3 Organise Resources 4 Implement Strategy Growth Strategy Branding Strategy Audit Committee Expectations Internal drivers Internal Insights Market Strategy Product Strategy Audit Strategy Executive Views Risk Views External insights Industry trends Operational Strategy Regulator Expectations External drivers Critical Strategic Requirements Audit Resources Co-source Resources Define • Agree Charter • Agree scorecard • Define structures Audit infrastructure (GRC, IDEA, EUC’s etc) Plan • Conduct risk assessment • Evaluate against each strategy • Develop audit plan (Targeting process) Audit Plan Execute • Prioritise work • Execute plan • Periodically recalibrate the plan Risk and business readiness assessments Evaluate • Review audit performance • Re-evaluate strategy and audit plan MI on key processes, critical controls, hot spots, and issues Improve • Continuously improve • Adapt to emerging business and audit requirements 4

Building an audit capability from scratch Universe Methodology People Systems and tools 5

Why use IDEA? n You ca data t r o p im size y n a f o its s e k a M o easy t m perfor of s e s y l ana data The source data is protected (read only) You can select samples using several sampling techniques Helps you develop tests which are repeatable n IDEA ca a in a t a d t r o imp e of g n a r e d wi s. file type sy to a e t i Makes les b a t join da n match ca u o y o s are p m o c or iles f t n e r diffe ally c i t a m Auto sa e t a r e gen ory t s i h e et compl uments c that do alysis the an 6

Small steps – our experience so far Supporting audit testing. Examples: • Identifying exceptional items • Performing analyses • Checking calculations • Cross-matching data between systems • Testing for gaps and duplicates • Sampling CASE 1: Credit Risk Acquisitions Decisioning Strategy Dummy data used to prove test design 7

Small steps – our experience so far Supporting audit testing. Examples: • Identifying exceptional items • Performing analyses • Checking calculations • Cross-matching data between systems • Testing for gaps and duplicates • Sampling CASE 2: Checking application of correct fees and charges to all Personal Current Accounts (PCA) and Credit cards • Fees and changes must be charged in accordance with published tariffs and aligned to T&C’s • High customer service and regulatory interest in accuracy • 1. 6 Million transactions (approx 80 GB) • Data transfer, cleansing and joins using IDEA • Initial analysis showed 99% accuracy • Deep dive on exceptions to understand differences • Positive assurance over whole population = value add 8

The journey continues - our migration to a better place Data Sources ETL Data Warehouse TSB will soon migrate to our own IT platform Opportunities to: • enhance audit access to data • develop our capabilities Business Intelligence 9

Developing a road map Knowing where you are is a good place to start. We have had our successes but are at an early stage of the data analytics journey. Our migration creates an opportunity to develop our approach and look at: • • Team structure and skills Technology and techniques Training Wider use of data analytics 10

Questions and answers 11