Building a BI and analytics data mart for a large UK building society

Services

Data Engineering, Data Enablement, Data Mart, Business Reporting

‍

Challenge

Whilst the client was data rich, the data ecosystem they were working within was highly inefficient. Multiple systems, locations for business logic, solutions built in isolation to satisfy short term needs. This created a largely ungoverned environment in which risk of error was high and usability low. Previous attempts to solve this had failed, often with very expensive big bang solutions. Projects would run out of funding/patience, and would be left half complete - compounding the issue.

‍

Proposal

Our focus was on pragmatism. Centralising existing robust data assets where viable, and sitting alongside newly created assets, built up piece by piece, in order to provide a customer-centric data model covering products, interactions and reference data.
 
Highly engaged users at each step, with all designs prototyped, and subject to user sign-off and acceptance into BAU operations.
 
Focusing always on delivering agnostic granular foundations with high reusability, on top of which business data assets are built to provide KPIs and/or analytical outputs. Always with a line of sight from summary to granularity.

Focus on moving business logic out of BI tools/systems and into the aforementioned layers, to minimise isolated logic.
 
Creation of high quality documentation to ensure high data literacy amongst the user communities, and to expose lineage, data definitions and policies.

‍

Team

COO, CTO, Programme Manager, Business Analysts, Data Modellers

‍

Outcome

Enabled users far quicker and easier access to fundamental datasets without needing to navigate several systems and centralise their own data within BI tools / SQL Server. Enabled decommissioning of multiple complex legacy reports and repointing to this centralised solution. Increased query speed and accuracy of several high-profile data entities (e.g. Current Account transactions, Mortgage application journeys). Provided increased granularity behind reporting, allowing 'so what?' analysis. Created an internal (to client) legacy, helping shape the teams, upskill the individuals, and breathe life into one of their almost redundant, expensive data platforms - setting them up to continue building up the data asset and looking after its relevance and accuracy once we have left the building.