The real symptom is disagreement, not effort
Manual effort is not, on its own, a reason to replace a system. Plenty of businesses run efficiently on simple tools because the volume does not justify anything more. The signal that actually matters is when two people can produce two different answers to the same question and both can defend them.
Once that happens, every meeting acquires a preamble about whose numbers are right, decisions get deferred until the figures are reconciled, and the reconciliation itself becomes somebody’s job. That is the point at which the systems have started costing more than they save.
Eight things that usually mean it is time
None of these is decisive on its own. Three or four together generally are.
- Somebody has to reconcile two systems before anybody trusts a number, and it happens more than once a month.
- The stock figure is treated as an estimate. Orders are promised on it and then checked physically before anybody commits.
- Nobody can say what a specific job or order actually made without an afternoon of work, so nobody asks.
- A key spreadsheet has one owner, and their annual leave is a business risk.
- The same information is keyed into two or more systems - order to accounts, accounts to stock, stock to a courier portal.
- Purchase commitment is first visible when the invoice arrives, so budget conversations are always retrospective.
- An audit, a customer questionnaire or a recall drill requires assembling evidence rather than printing it.
- Growth is being turned down, or delayed, because the admin behind it would not cope.
If you recognise fewer than three of these, you are probably fine for now. Fixing one specific process well beats buying a platform you are not ready to use.
Why the accounts package cannot fill the gap
The most common intermediate step is to push more work into Xero, Sage or QuickBooks - using their inventory features, adding an app for stock, adding another for quoting. It works up to a point, and the point is usually reached faster than people expect.
Accounting packages are built around the transaction: the invoice, the bill, the payment. They are excellent at that. What they do not model is the work that happens before the transaction - the enquiry, the quote, the specification, the works order, the route through the shop floor, the pick, the pack, the delivery. Those are the things that determine whether the invoice is profitable, and they are exactly what an accounting package is not designed to hold.
This is not an argument for replacing your accounts package. Most of our customers keep theirs, and the operational system posts into it.
What waiting actually costs
The cost of staying put is real but diffuse, which is why it gets discounted. It shows up as stock written off because nobody could see it was ageing, jobs quoted from an estimate that has been wrong for two years, parts fitted on site and never invoiced, rebates unclaimed, and orders promised on dates that were never achievable.
Most businesses find that when they finally measure one of these, it is larger than the annual cost of the system they were hesitating over. That is not a sales argument - it is an argument for measuring one of them before you decide either way.
What to do before you buy anything
- 1Pick the single question that causes the most argument - job profitability, stock accuracy, delivery performance - and measure it properly once, by hand if necessary.
- 2Write down where data is currently keyed twice. The list is usually longer than anybody expects and it makes the case on its own.
- 3Identify which spreadsheet would hurt most if its owner left tomorrow. That is your real risk, and it is worth addressing regardless of what you buy.
- 4Agree internally what "better" would look like in numbers, not adjectives. "Fewer errors" is not a target; "stock accuracy above 98%" is.
- 5Only then talk to suppliers, and lead with those numbers rather than asking for a demonstration.