First, the defence
Most of your spreadsheets are fine, and that matters
We do not have an office in Saudi Arabia
We serve organisations there from our Riyadh office and our engineering hubs in Lahore and Karachi, travelling for the parts of an engagement that need to happen in person.
A consultancy that tells you to replace every spreadsheet is selling software, not judgement. The spreadsheet is the most successful end-user computing tool ever made: a manager models a scenario, an analyst explores a data set, a team plans a quarter — all without a project, a budget or a queue for IT. An organisation that took that away would be slower and would deserve the resentment it earned. The starting position of any honest review is that the overwhelming majority of workbooks should be left exactly where they are.
The failure mode is narrow and specific. Somewhere in the estate, one workbook stopped being a personal tool and became shared infrastructure: several people edit it, a revenue-bearing or payroll-bearing process depends on it, and its formulas encode rules the business could not restate from memory. Nothing announced the crossing — it happened one convenient shortcut at a time. The work is to find those few files, name what they have become, and deal with them deliberately, while leaving everything else alone.
The crossing
Six signals a file has crossed from tool to infrastructure
One of these alone is survivable. Two or more on the same workbook is the definition of a system running without any of a system’s protections.
More than one pair of hands
Several people edit it, versions fork into final_v7, and an overwrite loses an afternoon’s work. Spreadsheets have no real concurrency; the copies are the symptom.
Money moves because of it
Pricing, commissions, payroll inputs or invoicing amounts come out of its cells. A formula error here is not a typo; it is a mispriced contract or a wrong salary.
Errors surface weeks later
A dragged formula or a broken reference produces plausible numbers, and the discovery comes from a customer or an auditor, not from the file. Silent wrongness is the spreadsheet’s characteristic failure.
Its author is its only maintainer
One person understands the tabs, the hidden columns and the order things must be done in. Their leave is a delay; their resignation is an incident.
Other files feed from it
Downstream workbooks link to its cells, so a structural change here breaks reports elsewhere — the surest sign it has become a data source pretending to be a document.
Someone official asked how a number was made
An auditor, a regulator or a board member wanted the derivation of a figure, and the answer was a tour of formulas. Provenance is precisely what a workbook cannot testify to.
The options
Three honest destinies for a load-bearing workbook
Replacement is the loudest option and often the wrong first one. Each destiny fits a different kind of workbook, and the fit is decided by blast radius, not by how untidy the file looks.
| Option | Fits when | What changes for the team | What it will not fix | Best for |
|---|---|---|---|---|
| Harden it in place | One team uses it, the blast radius is contained, and the logic is sound — the risk is operational hygiene, not the tool. | Very little: input validation, locked logic cells, a named owner, one master copy and a documented monthly routine. | Concurrency and audit. Ten editors and a provenance question will still defeat it. | Single-team tools whose failure inconveniences a department, not the business. |
| Move the data, keep the habit | The workbook is really a database plus analysis: the risk is in storage and feeding other reports, while the analysis itself is legitimate spreadsheet work. | The data lives in governed warehouse tables with history and access control; the spreadsheet becomes a live view of it, and the analysis habits survive intact. | A broken multi-user process. If the workflow itself is the problem, relocating its data does not mend it. | Reporting and analysis workbooks that quietly became a data source for others. |
| Replace the workflow with software | Many users, money at stake, rules that must be enforced rather than trusted — the conditions under which formulas are the wrong material. | A small internal application with permissions, validation, history and an audit trail — scoped to the one workflow, not an enterprise platform programme. | The speed of ad-hoc change. A formula edit took minutes; software changes take a release, and that trade must be entered knowingly. | Multi-user, revenue-bearing processes where a silent error costs real money. |
Where we fit
Triage before construction, and the case for leaving files alone
- 01
The inventory, ranked by blast radius
We find the workbooks the business actually depends on — following the money and the month-end, not the file count — and rank them by what breaks if each is wrong for a week. Most of the estate ranks nowhere, and we say so.
- 02
A recommendation that includes doing nothing
Each ranked workbook gets one of the three destinies with the reasoning written down — and where the honest answer is leave it, protected by an owner and a routine, that is the recommendation you get.
- 03
The formulas treated as the requirements
When a workbook is replaced or migrated, its formulas and hidden rules are read as the specification of what the business actually does, so nothing the sheet quietly handled is lost in the new system’s first month.
Excel stays on every desk for the exploration and modelling it is genuinely best at; what moves out of it is the storage, the shared rules and the money — the parts a spreadsheet was never built to guard.
Common questions
What buyers ask
If we replace the big workbook, does the team lose Excel?+
No, and a project framed as taking Excel away deserves the resistance it will meet. What moves is the storage of shared data and the enforcement of shared rules; what stays is everything Excel is actually good at — exploration, one-off analysis, modelling a scenario before it becomes policy. In the migration destiny the team often keeps working in spreadsheets that now read live from governed tables, so the daily habit barely changes while the fragile part disappears underneath it.
The workbook has run for years without an incident. Why touch it now?+
Because a spreadsheet’s characteristic failure is not a crash you would have noticed — it is a plausible wrong number, produced silently and acted on. Years without a visible incident is not evidence of correctness; nobody is checking, which is the condition itself. The second reason is the author: the file is one resignation away from being unmaintainable, and that risk grows while everything looks calm. Neither argument means replace it — hardening in place may be the whole answer — but both mean the file deserves a deliberate decision rather than continued luck.
How do we work out which files the business truly depends on?+
Not by scanning the file server — by following two threads. Follow the money: ask which files a price, a commission, a payroll input or an invoice passes through on its way to being real. And follow the month-end: ask which files, if locked for three days, would delay the close or a customer-facing commitment. The people who know are the operators, not the org chart, and the honest question to ask them is which file they would least like to open and find corrupted. The list that produces is short, and it is the whole inventory that matters.
In this cluster
An inventory session that ranks the load-bearing workbooks by blast radius and ends with one agreed candidate for hardening, migration or replacement.
Show us the workbook
