Practice

Business Intelligence

Reporting, dashboards and management information rebuilt as governed BI data products — consistent KPIs across functions, traceable to source, owned by the business.

Talk to our BI team →
Business Intelligence A "one number, three audiences" diagram. The same KPI — capital ratio 14.8% — is shown on three executive dashboards (board, finance, risk) all sourced from a single governed data product, with a single accent marker linking the three views back to one definition. GOVERNED DATA PRODUCT capital_ratio_v.2026.04 · one definition · auditable BOARD CAPITAL RATIO 14.8% ↑ 30bps FINANCE CAPITAL RATIO 14.8% ↑ 30bps RISK CAPITAL RATIO 14.8% ↑ 30bps ONE NUMBER · THREE AUDIENCES · ONE DEFINITION Business Intelligence A "one number, three audiences" diagram. The same KPI — capital ratio 14.8% — is shown on three executive dashboards (board, finance, risk) all sourced from a single governed data product, with a single accent marker linking the three views back to one definition. GOVERNED DATA PRODUCT capital_ratio_v.2026.04 · one definition · auditable BOARD CAPITAL RATIO 14.8% ↑ 30bps FINANCE CAPITAL RATIO 14.8% ↑ 30bps RISK CAPITAL RATIO 14.8% ↑ 30bps ONE NUMBER · THREE AUDIENCES · ONE DEFINITION

Practice signals 1 / 3

Definitions before dashboards

BI succeeds when teams agree the metric before they admire the visual.

Who this is for

Heads of BI, heads of finance MI, executive analytics leads, CDOs and finance transformation owners at insurers, banks and asset managers.

Where we help

Most BI estates have grown by accretion. Each function builds its own dashboards over its own extracts. The same KPI ends up calculated three different ways, in three different tools, with three different lineages. Reconciling them at executive level becomes a monthly negotiation. Nobody trusts the dashboard they did not build.

What we do

  • Data taxonomy and definitions. Defining the key financial metrics, value drivers and reference data that underpin every KPI — so the same number means the same thing in every report.
  • Balanced scorecard design and rollout. Key indicators weighted by importance, sourced from the business units that own them, with the data pipeline and quality indicators behind them. We have led these end-to-end at a large South African insurer.
  • Self-service BI rollout. Decommissioning legacy reporting platforms (we have led a private-wealth BI estate replacement end-to-end) and replacing them with self-service environments executive teams actually use.
  • Performance reporting. Time-weighted rate of return, assets under management, attribution and transaction reporting — production reporting, not slideware.
  • IFRS 17 ETL and reporting. Balance sheet, income statement, annual headline earnings, KPI and profit-analysis reports — built on the calculation engine, not bolted on top.
  • BI for risk and fraud. Predictive analytics for claims-fraud hotspot detection, KRI reporting, risk control architecture — BI applied to risk, not just finance.
  • Re-engineer existing BI estates so dashboards consume from a governed data product, not a spreadsheet extract.
  • Reference and master data management. Code and description repositories, conforming codes across business units, data lineage and quality monitoring.
  • Versioned KPI catalogue. One number, one definition, one source — business-owned, not buried in someone’s workbook.
From dashboard sprawl to governed BI data products A before/after comparison. Before: each function has its own dashboard, calculating the same KPI three different ways from three different extracts, producing three different numbers. After: each function still has its own dashboard, but they all consume from one governed BI data product, so the number agrees. BI ESTATE — BEFORE / AFTER From dashboard sprawl to one governed number BEFORE — EVERYONE COMPUTES THEIR OWN FINANCE 14.8% RISK 14.6% EXEC 15.1% FIN extract .xlsx · monthly RISK feed .csv · weekly Exec brief slide · ad-hoc Three numbers, three lineages, one monthly argument. Reconciliation happens after the meeting. Trust falls. AFTER — ONE GOVERNED DATA PRODUCT FINANCE 14.8% RISK 14.8% EXEC 14.8% GOVERNED BI DATA PRODUCT capital_ratio_v.2026.04 · versioned · auditable One number, three audiences, one definition. Reconciliation built into the product from the start. From dashboard sprawl to governed BI data products A before/after comparison. Before: each function has its own dashboard, calculating the same KPI three different ways from three different extracts, producing three different numbers. After: each function still has its own dashboard, but they all consume from one governed BI data product, so the number agrees. BI ESTATE — BEFORE / AFTER From dashboard sprawl to one governed number BEFORE — EVERYONE COMPUTES THEIR OWN FINANCE 14.8% RISK 14.6% EXEC 15.1% FIN extract .xlsx · monthly RISK feed .csv · weekly Exec brief slide · ad-hoc Three numbers, three lineages, one monthly argument. Reconciliation happens after the meeting. Trust falls. AFTER — ONE GOVERNED DATA PRODUCT FINANCE 14.8% RISK 14.8% EXEC 14.8% GOVERNED BI DATA PRODUCT capital_ratio_v.2026.04 · versioned · auditable One number, three audiences, one definition. Reconciliation built into the product from the start.

Diagram signals 1 / 3

Definitions before dashboards

BI succeeds when teams agree the metric before they admire the visual. The body diagram should make this route explicit enough to discuss in a working session.

From dashboard sprawl to one governed BI data product.

Outcomes

  • One executive number per KPI, traceable end-to-end.
  • Reduced dashboard sprawl and tool duplication.
  • Faster onboarding of new metrics, with controls.
  • BI that survives a leadership change, because it is documented and owned.

Engagement model

We typically start with a BI Assessment of a single executive pack (board, ALCO, ExCo or operations) and the data products feeding it — then expand into a governed, business-owned BI estate.

Related software

Symphony ReportingSymphony Automate

Talk to our BI team

Book a 30-minute conversation See diagnostic assessments Read our case studies