Work / Case

Turning data quality concerns into measurable action

8,000+

issues identified and resolved before go-live

Role: Product Owner (consultant)
Context: Data migration to a cloud HR platform at a large Nordic transport and logistics company
Timeframe: 2022–2024

Problem: Suspected data quality issues, scale unknown
Contribution: Owned a Power BI data quality solution
Outcome: 8,000+ issues identified and resolved before go-live
Limitation: Root causes not tracked over time

AI-generated illustration of a review meeting: four people around a conference table looking at a wall-mounted screen showing an HR data quality dashboard, with a handwritten go-live checklist on a glass whiteboard to the left — AI generated illustrationReviewing the open list with the people who owned it.

The situation

A system migration, and a concern nobody could size.

The company was migrating its HR data from a legacy system to a cloud HR platform. Leadership was concerned about the quality of the existing data, but there was no clear picture of how many issues existed, where they were, or what needed to be fixed before go-live.

Making the problem visible

Measure what had previously been a gut feeling.

As Product Owner, I owned a Power BI solution that monitored HR data quality and gave stakeholders a shared view of the issues and their scale.

The analysis surfaced more than 10,000 data quality issues. With the problem now visible, we could move from general concern to prioritising what needed attention first.

Illustration: a dense tangled cloud of many coloured points joined by chaotic curved lines on the left passes through four upright translucent panes, one colour each, and emerges on the right as four neat, evenly spaced horizontal rows of dots — Thousands of issues, filtered down to a list that could be worked through. Illustration, not a screenshot.

From measurement to action

Prioritise the issues with the most impact, then stop them coming back.

I worked with HR business partners and other stakeholders to identify and prioritise the issues with the most impact, then drove the remediation work with the client’s HR shared services team.

Beyond the immediate fixes, we established a process to help prevent the same issues from returning. The goal was not just to clean up the data for migration, but to maintain its quality going forward.

Illustration of the data quality process in three steps: a DQ engine reading HR systems, an errors report, and a remediation team of three people, with two return paths back into the systems labelled Correct and Implement — The process: detect, report, fix, then feed the fix back into the systems. Illustration, not a screenshot.

The outcome

Over 8,000 issues identified and resolved before go-live.

More than 8,000 issues were identified and resolved ahead of go-live. Payroll ran without problems after the migration, and a process was in place to manage data quality beyond the project.

05 / What I’d do differently today

Fixing an error is one thing. Making sure it doesn’t happen again is another.

I’d pay closer attention to whether the underlying causes were actually addressed. Today, I’d track recurring error types over time and use that evidence to identify which processes needed to change.

Next case

Rescue vest, bachelor thesis, 2018

When the user can't think straight, remove the wrong choice

Read the case

If this sounds familiar, let’s find out.

Two to four weeks from uncertainty to a decision you can act on.

Let's explore together