Kristine Walker

Case studies / 03

Mock case study · 03

Open Access, Certified Answers

A data governance approach for a post-acquisition insurer where three teams bring three different numbers

Type
Mock case study — fictional company, systems and figures, built to show approach and artifacts
Company
Lone Star Mutual (fictional), a regional auto and home insurer that has just acquired a smaller carrier
My role
Senior Business Analyst, data governance
Grounded in
Real experience tracing the data behind 300+ executive metrics at a large financial services firm after a major acquisition
12core metrics, one certified definition each
3competing loss ratios — all labeled
Monthlymetric check-ins
100%target: certified or labeled numbers

1The problem

After acquiring a smaller carrier, Lone Star Mutual has two policy systems, open access to data, and analysts writing their own SQL. Every monthly executive meeting starts the same way: Underwriting, Finance and the acquired team each bring a different number for the same metric, and the first half hour goes to deciding whose is right.

Diagram of three teams producing three different numbers for the same metric
“I can tell you which table a number came from. I can’t tell you which number is the right one.”— BI lead (fictional)
“I don’t mind definitions changing. I mind finding out at audit time.”— Compliance Officer (fictional)

2The decision rule

Leadership doesn’t want to restrict access — it keeps the business fast. So the answer isn’t locking data down; it’s certifying the answers. The data owner and the governance committee set the standard, anchored to the compliance requirement. When they agree, that becomes the certified definition. A team with a legitimate different need keeps it as a labeled variant.

Decision rule for certifying a metric definition

Why three loss ratios can all be right

Version (fictional quarter, $50M earned premium)ResultLabel
Paid losses ÷ earned premium52.0%Variant — Claims (cash view)
Incurred losses ÷ earned premium61.0%Certified — Actuarial, statutory basis
(Incurred losses + LAE) ÷ earned premium71.0%Variant — Finance (includes claim-handling cost)

3The future state

Future-state governance cycle with certified metrics
NeedBeforeAfter
Which number is rightArgued in the meetingThe certified definition, by rule
OwnershipUnclear after the acquisitionNamed owner and steward per metric
Different team needsCompeting numbersLabeled variants
ChangesDiscovered at audit timeRequested, assessed monthly, logged
AccessOpenStill open — the outputs are certified instead

4The artifact: a working metric catalog

The core deliverable is an Excel workbook. For each of 12 insurance metrics it records the business question, certified definition, formula, owner, steward, source systems, refresh, compliance anchor, approved variants, status and next check-in. A live scorecard counts certified metrics and flags the next check-in automatically.

Workbook tabWhat it shows
Metric Catalog12 metrics, one certified definition each, variants listed separately
Three Versions of TruthHow each team defined policies in force, retention and loss ratio — and how each was resolved
Loss Ratio ExampleLive formulas: paid, incurred and loss-and-LAE ratios from the same inputs
Ownership (RACI)Who is responsible, accountable, consulted and informed
Change LogRequests, reasons, decisions and effective dates
Scorecard & GlossaryLive governance measures and plain-language terms
Phased governance roadmap

5What I learned

Metrics evolve, they don’t drift. The goal isn’t to freeze them, but to make sure every change is assessed and captured before audit time.
Lineage isn’t meaning. Tracing report code back to cloud data shows where a number comes from, not how it’s used or which version is right.
Certify, don’t restrict. Open access and trustworthy numbers can live together when the rules for what’s official are clear.

Lone Star Mutual, its people, systems and figures are fictional.

Downloads

The full case study includes requirements, stakeholder analysis and trade-offs.