An owner of a twenty-eight-person services firm opened her operations spreadsheet to me last month. Thirty-seven tabs. Six people editing it on different days of the week. Two people who genuinely understood how the cross-tab formulas worked, one of whom was due to retire in March. And at least one column whose formula nobody could remember writing, kept in place because removing it broke the monthly report. The spreadsheet had served the firm well for nine years. It was now actively slowing the business down, and she had not quite registered why.
That spreadsheet is doing a job it was never designed to do. It started as a free, fast, flexible answer at five people, which is exactly the right answer at five people. Somewhere between twenty and thirty-five staff, the same flexibility that made it useful becomes the structural liability that holds the firm back. The transition to a dedicated system is rarely planned and almost always happens too late. The signs are recognisable in advance, and the move is much less disruptive when made deliberately.
What does it mean for an SME to outgrow its spreadsheet?
Outgrowing the spreadsheet means the operational data inside it is now serving more people, more processes, and more decisions than its structure can carry safely. The flexibility that made it useful at five people becomes the source of the problem at twenty-five. The file has quietly become the system of record for what the whole firm depends on, and it lacks the controls a system of record needs.
Repeated audits of spreadsheets in active business use find that 88 to 94 per cent contain at least one error, and developer studies put cell-level error rates at 1.1 to 5.6 per cent in spreadsheets reported as complete. Those numbers compound through any calculation chain. At small scale the errors are visible to the person using the file. At twenty-five people, the same errors propagate into invoices, forecasts, and decisions before anyone notices.
What are the four signs the spreadsheet has become the bottleneck?
Four operational symptoms flag readiness in advance. The first is edit conflict and version proliferation: people make local copies, email files with “FINAL” appended, and discover their changes have been overwritten by someone else’s earlier version. The second is formula integrity: cells contain values nobody can trace, formulas break for reasons nobody can explain, and the audit tools in Excel sit unused because they require literacy in spreadsheet auditing.
The third sign is key-person dependency. The structure of the file, why tabs are organised a particular way, what a column actually represents, which cells contain formulas rather than data entry, lives in one person’s head. If that person is ill, on holiday, or leaves, critical processes pause or proceed on guesswork. The fourth sign is the absence of an audit trail. Excel’s Track Changes does not meet the audit requirements expected under UK data governance frameworks or sector-specific compliance regimes, and a business that begins to operate in a regulated environment finds the gap fast.
Where will you actually meet these signs in practice?
You meet them in three places, usually in this order. First in the monthly close, where reconciling figures across people takes longer each quarter. Second when a key person is unavailable and a routine process stalls because the spreadsheet’s logic lives in their head. Third when a new starter cannot use the file without an hour of one-to-one teaching from the original author.
A fourth meeting point is the AI conversation. Owners often discover the bottleneck when they try to point an AI tool at the spreadsheet to automate a piece of it. The AI exposes the structural problems immediately, formula references that point at deleted cells, columns whose definitions have drifted over time, dates stored as text strings that look like dates. The tool is reading the file accurately and flagging what the spreadsheet was always hiding behind manual interpretation.
When should the migration happen, and how is it sequenced?
The right time is when at least two of the four signs are clearly present and the firm has six to twelve months of operational headroom to plan deliberately. Crisis migrations, where the spreadsheet has just failed and forced the conversation, cost two to three times more. Realistic implementation timelines for CRM and similar systems sit at two to four months once planning, data preparation, and user adoption are properly accounted for.
A proportionate sequence runs over twelve weeks in three phases. Weeks one and two are assessment: document the workflows the spreadsheet currently supports, identify duplicates and inconsistencies, run a migration simulation on a subset of data to confirm the new system can hold the logic. Weeks three to six migrate the single most isolated workflow, the client contact database for a services firm, time tracking for a professional services firm, inventory for a product business, and run that workflow in parallel with the spreadsheet for thirty days. Weeks seven to twelve add the second and third workflows, with parallel running maintained until each replacement has demonstrated reliable operation for sixty to ninety days in production conditions.
What deliberately stays in a spreadsheet, even after the migration?
Some workflows are legitimately spreadsheet-shaped, and forcing them into a system is a mistake. Exploratory analysis is the clearest example. When a manager wants to test a hypothesis or build a one-off forecast, a spreadsheet provides rapid iteration that a formal report request does not. Capacity planning and resource scheduling often belong in spreadsheets too, because the structure of the question changes month to month.
What changes is governance. The spreadsheets that remain in use do so with explicit recognition of their role. They sit alongside the system of record without replacing it, they carry no compliance or audit weight, and they exist as tools for thinking rather than as infrastructure for running the firm. That distinction is the difference between an owner-operated business that has grown into systems maturity and one that has tried to impose systems thinking on inherently spreadsheet-appropriate tasks. The mature stance is to keep both, and to know which is which.
The owners who recognise the bottleneck early and plan the move deliberately spend less in total than the owners who wait. Crisis migrations cost two to three times more in operational disruption, in remediation, and in the external help required to fix problems that better preparation would have prevented. The signs are visible months in advance, and the proportionate work fits inside a single quarter. If you want help working out which workflow to move first and how to sequence the rest, book a conversation.
