When your spreadsheet becomes a business risk

Almost every business we work with started on spreadsheets, and they were the right choice at the time. Excel is fast, flexible and free. The problem is that spreadsheets do not announce when they stop being adequate - they just quietly become the single largest operational risk in the company.

Five signals it is time

1. 1One person understands the file. If someone leaving would create a crisis, that is not a tool, it is a dependency.
2. 2The numbers disagree. Finance says one stock figure, the warehouse says another, and reconciling them is a weekly ritual.
3. 3You cannot see today. Reporting requires someone to build the report, so decisions are made on last month's picture.
4. 4Data gets rekeyed. The same information entered into three places is three chances to be wrong.
5. 5There is no history. You cannot establish who changed a figure, when, or what it was before.

Do not build everything at once

The classic ERP failure is trying to replace all operations simultaneously. It takes eighteen months, costs three times the estimate, and the business changes underneath it before it lands. Sequence it instead:

1. 1Start with the department in the most pain - usually inventory or sales. One module, in production, in eight to twelve weeks.
2. 2Run it in parallel with the spreadsheet for a few weeks until the numbers reconcile and staff trust it.
3. 3Add the adjacent module that shares data with the first. Each addition compounds the value of what is already there.
4. 4Integrate finance last. It is the least tolerant of instability and benefits most from clean upstream data.

Off-the-shelf or custom?

Buy when your process is genuinely standard - straightforward accounting, payroll, common retail. Odoo, Zoho and Sage are mature and far cheaper than a build.

Build when your process is your advantage. If you have a pricing model, a credit arrangement or a logistics approach that competitors do not, forcing it into generic software throws away the thing that makes you money. We have also seen the reverse: a client convinced they needed custom, whose process turned out to be entirely standard. We told them to buy Odoo. That is a better outcome than a build they did not need.

Budget for the parts nobody quotes

• Data migration and cleaning - often the single most underestimated line
• Training, including the staff who will resist it
• A parallel-running period where both systems are maintained
• Three to six months of adjustments as reality meets the design

A realistic figure for a first module in Kenya is KES 650,000 to 1.5 million. Against two staff spending half their week reconciling numbers, the payback is usually under a year - and the risk that leaves the building with the spreadsheet owner does not show up on any invoice.