When a spreadsheet stops being enough.

Spreadsheets are better than people admit

Let's start on the right side. A spreadsheet is free, instantly editable by anyone, and understood by everybody who's ever worked in an office. For a business planning a handful of jobs a week, it is genuinely the correct tool, and anyone telling you otherwise is selling something.

Most of the "digital transformation" pitched at small manufacturers is a solution looking for a problem. So the useful question isn't whether spreadsheets are unprofessional. It's whether yours has started costing more than it saves.

The five signs it's costing you

1. Somebody has to be the human database

There's one person who knows which tab is current, why row 47 is highlighted amber, and which of the three files is the real one. When they're on holiday, planning slows down. That person isn't a bottleneck because they're bad at their job — they're a bottleneck because the spreadsheet needs a human interpreter.

2. Two people can't safely use it at once

The moment planning requires coordination about who has the file open, the tool is shaping the business rather than the other way round. Cloud spreadsheets soften this, but the underlying problem remains: nothing stops two people entering contradictory information, and nothing tells you afterwards which one was right.

3. Mistakes are only found downstream

A job gets scheduled into a week where its materials haven't arrived. Nothing objects — a spreadsheet will cheerfully accept any date you type. The error surfaces on the shop floor a fortnight later, when fixing it means moving three other jobs. This is the expensive one, and it's the clearest signal, because software can check rules and a grid of cells cannot.

4. Reporting means an afternoon of copying and pasting

If answering "how did last month go?" or "what's our capacity in October?" takes someone half a day of manual assembly, you're paying that cost every single month, forever. Most people never add it up. A half-day a month at a manager's cost is several thousand pounds a year.

5. Nobody trusts it enough to rely on it

The clearest sign of all: the spreadsheet exists, and people still ring each other to check. When the official system is routinely verified by informal conversation, the official system has already failed — it just hasn't been replaced yet.

What a first tool should actually do

The classic mistake is trying to replace everything at once. A twelve-month build covering quoting, planning, stock, invoicing and reporting will be late, expensive and rejected by the people meant to use it.

A first internal tool should cover one workflow, and should be the one that hurts most. Usually that's planning — because planning errors are the ones that cost money downstream. Three things make the difference between a tool that gets used and one that gets abandoned:

• It must be faster than the spreadsheet for the daily task. If entering a job takes longer than it used to, people will keep a private spreadsheet alongside it, and you'll have two systems instead of one.
• It must catch the mistakes the spreadsheet let through. That's the entire reason for building it. Deadline conflicts, impossible sequences, capacity that doesn't exist — the software should object at the point of entry.
• It must survive a closed laptop. Real database, real accounts, automatic saving. Nobody should lose an afternoon's planning because a tab was shut.

A real example

David Jackson & Son make bespoke timber windows and doors in Shropshire. Every order is a different shape, and each one moves through eight production stages. That was planned on paper and in people's heads.

What replaced it was deliberately narrow: an eight-stage drag-and-drop scheduler that plans jobs across the workshop and automatically flags a breach when a stage would overrun its deadline. Not quoting. Not invoicing. Not stock. One workflow, built properly.

Under it: a Supabase database so the plan is shared rather than trapped in one browser, two-factor authentication because the plan carries customer names, autosave on every change with one-click backup and restore, and a GitHub-to-Vercel pipeline so every change is reviewable and reversible.

It's worth noting how it started — on browser storage, to prove the workflow was right before spending anything on infrastructure. It earned the database by being used. That order matters: spend follows evidence, not hope.

The full case study is here, including a screenshot of the planner in use.

What it costs

This is where most businesses get frightened off, usually by an agency quote. UK agencies commonly quote £20,000–£80,000 for an SME software project, and for a business turning over a few hundred thousand that's simply not happening.

But that band isn't the only option:

• A small internal tool covering a single workflow — a scheduler, a planner, a tracker — is £6,000–£10,000.
• A business application with multiple connected workflows, a database, admin and authentication is £10,000–£25,000.

Running costs afterwards are usually a few pounds a month on hosting, in accounts in your own name. And it should come with handover documentation written so another developer could take it on — being irreplaceable is a business model I'd rather not have.

Before you commission anything

The single most common reason internal tools fail is that they encode how a manager thinks the work happens rather than how it actually happens. The only reliable fix is watching the real process, on site, before anything is designed.

That's what the Groundwork Visit is: I come to your premises, map the workflow as observed rather than as described, and write you a scope with a fixed price. If the honest conclusion is that you need a better spreadsheet rather than a system, the document says so — and you'll have spent a few hundred pounds to avoid spending thousands.