From spreadsheets to visibility: building better reporting systems

Spreadsheets are not the enemy. They are the most flexible analytical tool most organisations will ever have, and any reporting programme that begins by treating them as a failure is starting from the wrong place.

The problem is not the spreadsheet. It is the spreadsheet doing a job it was never designed for: acting as the system of record, quietly, for something the whole business depends on.

How you know you have crossed the line

• A file has an owner, and when that person is away the number cannot be produced
• There are versions with names like final-v3-updated-new
• Someone rebuilds the same workbook every month from the same exports
• Two departments quote different figures for the same metric and both can defend theirs
• Nobody can explain how a total is derived without opening the file and tracing formulas

At that point the spreadsheet has stopped being analysis and become infrastructure — without any of the controls infrastructure normally gets.

What to move, and what to leave

The instinct is to replace everything. That is usually wrong, and it is why so many reporting projects overrun.

Move the things that must be consistent, auditable, and available without a particular person: transactional records, balances, stock positions, anything that feeds a statutory or board report. Leave the things that benefit from being fast and disposable: scenario modelling, one-off analysis, a quick view someone needs this afternoon and will never open again.

A good reporting system does not eliminate spreadsheets. It removes the ones that were doing a system's job, so the rest can go back to being useful.

Build the layers in the right order

Reporting maturity has a sequence, and skipping a step reliably produces a dashboard nobody trusts.

• Capture — the operational event is recorded once, at source, by the person doing the work
• Reconciliation — the system can prove its own numbers tie out, without a human assembling the proof
• Reporting — standard views that are the same for everyone, derived from the reconciled data
• Analysis — flexible exploration on top, where spreadsheets legitimately return

Most failed reporting projects start at step three. They build attractive dashboards on data that was never reconciled, users find a discrepancy in the first week, and trust never recovers. Rebuilding trust in a number costs far more than getting it right initially.

What good looks like

You know reporting is working when arguments change character. Instead of debating whose figure is right, people debate what to do about a figure everyone accepts. That is the entire return on the investment.

The dashboards are the visible part. The reconciliation underneath is the part that makes them worth looking at.