Edcen Data All articles
Regional Business

The Knowledge in One Head: Testing Whether Your Data Strategy Survives Its Builder

Edcen Data
The Knowledge in One Head: Testing Whether Your Data Strategy Survives Its Builder

Photo by Photo by Flipsnack on Unsplash on Unsplash

There is a particular kind of quiet in a regional office the week after the analyst resigns. The dashboards still render. The monthly pack still exports. But somewhere in the middle of it, a filter that nobody else understands, applied to a table that nobody else can locate, has stopped being maintained — and the numbers will be subtly wrong for a fortnight before anyone notices.

Key-person risk in data teams rarely announces itself. It looks like competence. The person who always knows which of the four customer tables is the real one is efficient, helpful and, from a governance standpoint, a single point of failure carrying a full reporting estate on their shoulders.

The bus factor, in plain English

Risk specialists talk about the bus factor: how many people would have to be unavailable before a function stops working. In data teams at small and mid-sized regional businesses, the honest answer is frequently one. That is not a criticism of the individual. It is a structural observation about how lean teams absorb complexity. When a firm of four hundred people has two analysts, the accumulated logic of a decade's reporting naturally pools in whichever of them has stayed longest.

What makes this more acute outside the major cities is straightforward. Teams are smaller, so overlap is thinner. Spans of responsibility are wider, so time for documentation is scarcer. Recruitment for specialist skills is slower and more expensive, which lengthens the gap between departure and replacement. And acquisitions of neighbouring businesses — a familiar pattern in regional professional services, manufacturing and distribution — often bring reporting estates together without bringing their definitions together.

The commercial consequence is not theoretical. Due diligence during a sale, an audit query that requires an explanation of a prior-year figure, or a regulatory return that needs to be reproduced to the same method as last time: each of these depends on knowledge that may exist in exactly one head.

Where the knowledge actually hides

The risk is rarely in the headline system. Look instead at the following:

None of these is unusual. All of them are recoverable, provided the organisation tests for them before it has to.

The handover test: five exercises

A handover test is not a document. It is a set of activities that demonstrate whether knowledge can move. Five are enough for a first pass, and each can be run inside a fortnight.

  1. The silent week. For five working days, the primary analyst answers no questions about reporting logic. Requests are logged instead. The size of the log is the size of the dependency.
  2. The rebuild. Ask a colleague unfamiliar with a Tier 1 report to reproduce it from the documentation alone. Every question they need to ask is a gap.
  3. The definition quiz. Give three managers the same core metric and compare the definitions they write down. Divergence between them is a governance gap even where no departure is imminent.
  4. The access audit. List everyone with administrative rights and everyone whose name appears on a scheduled job or integration. Anything tied to a personal account is a finding.
  5. The departure drill. Walk through a leaver's final week on paper. What is transferred, in what order, and who signs it off? If the answer is "they'll do a handover", there is no process.

A checklist that fits on one page

Documentation is the usual recommendation and the least useful on its own. Four dimensions need to be covered, and each needs a named owner.

Dimension What to record Test of adequacy
Documentation Definitions, source of record, lineage, known limitations A colleague can rebuild the output without asking a question
Ownership Named owner per metric and per report, with a deputy Every Tier 1 output has two names attached
Access Role-based permissions, service accounts, vendor contacts No scheduled job or licence sits with an individual
Decision rights Who can change a definition, who approves, who is informed A change can be made and evidenced without a meeting

Decision rights are the most commonly omitted and the most consequential. Firms frequently document how a number is calculated while leaving entirely unclear who is entitled to change it. The result is drift: small, well-intentioned adjustments that accumulate into a series nobody can reconcile across years.

Resilience, not another dashboard

There is a reflex in UK data strategy discussions that treats maturity as a matter of output: more dashboards, more sources, more self-service. Output is easy to demonstrate and easy to fund. Resilience is harder to show on a slide, and it does not photograph well.

But resilience is the better measure. A strategy that depends on one person's memory is not a strategy; it is a personal capability with a corporate budget attached. The firms that handle departures, audits and acquisitions calmly are rarely the ones with the most sophisticated tooling. They are the ones where a number can be explained by two people, where access is not granted by habit, and where the right to change a definition is written down.

Questions worth putting to the board

Three questions will surface most of the exposure without a consultancy engagement. If our lead analyst left on Friday, which reports would we be unable to rebuild by the end of the month? Which of our core metrics has a definition that two managers would write differently? And who, by name, is permitted to change how a reported figure is calculated?

Answers that take more than a few minutes to assemble are the finding. The fix is not a new platform. It is a handover discipline, applied while the person who holds the knowledge is still there to disagree with the documentation — which is the only reliable way to know that it is right.

All Articles

Related Articles

Rungs and Regions: A Data Maturity Checklist for UK Organisations

Rungs and Regions: A Data Maturity Checklist for UK Organisations

Beyond Gut Feel: A Finance-Proof Way to Value Data Work in 2026

Beyond Gut Feel: A Finance-Proof Way to Value Data Work in 2026

When More Tools Mean Less Insight: A 2026 Right-Sizing Audit for UK Firms

When More Tools Mean Less Insight: A 2026 Right-Sizing Audit for UK Firms