Case study · Insurance

460 dashboards to 72, licence spend cut 38%.

A Lloyd's-market specialty insurer had a BI estate spread across three platforms and 460 dashboards — most of them unused or duplicated. A six-month rationalisation programme rebuilt the top 72 on a governed semantic layer.

460 → 72
Dashboards
−38%
Licence spend
2.1×
Adoption
6 mo
End-to-end
Overview

Noise out, signal in.

Usage telemetry revealed that 83% of dashboards had fewer than 3 users per month. Six metrics had different definitions in different dashboards. Analysts were re-pulling the same data weekly. We fixed both — by retirement and by semantic-layer discipline.

Approach.

01

Usage audit

Telemetry across Power BI, Tableau, Qlik. Evidence-based retirement list.

02

Semantic layer

dbt-based metric layer with owned, version-controlled definitions.

03

Top-72 rebuild

Audience-first dashboards with usage tracking from day one.

04

Self-service enablement

Curated datasets, office hours, guild.

05

Retirement

Communicated retirement of the other 388 dashboards, with owner sign-off.

−38%
Annual licence spend
£1.8m
Annual saving
2.1×
Active users
−67%
Time-to-insight
1
Source of truth per metric
72
Governed dashboards
Technology

Tools & frameworks we use.

Power BI
dbt
Snowflake
Azure DevOps
Monte Carlo
Cube
Ready when you are

Put your data to work.

Book a free 30-minute consultation with a senior Databuzz consultant.