← All posts Patterns we see

The spreadsheet that became a system when no one was looking, how a workbook turns into critical infrastructure

Almost every team has one workbook everyone is quietly afraid to touch. Here is how it got that way, and what to do about it.

Published 2026-06-164 min read
Tools we live in
n8n
Make
Zapier
Supabase
PostgreSQL
Cloudflare
HubSpot
Stripe
Shopify
Airtable
Slack
Notion

There is a spreadsheet in your business that everyone is a little afraid to touch. You probably know exactly which one as you read this.

How it happened

It started innocently. Someone needed to track a few things, so they opened a workbook. It worked, so they added a tab. Then a formula that referenced the first tab. Then a second person started using it, then a third, and at some point a process that the whole team depends on quietly moved inside a grid of cells that one person understands and nobody documented.

That is the pattern. A spreadsheet does not announce the day it stops being a spreadsheet. It just gradually takes on more weight until it is holding up part of the business, and the moment you notice is usually the moment something goes wrong.

The tells

The tells are consistent. One wrong paste and a column of numbers is gone, with no version history to recover them. A single person really understands how the formulas connect, and when they are on leave the work stops or, worse, carries on with mistakes nobody catches. New starters are told to be careful with it rather than shown how it works, because no one fully remembers anymore.

None of this means spreadsheets are bad. They are a brilliant place to start, precisely because they are so easy to begin. The problem is not the tool. It is that a tool meant for quick, personal calculation has been promoted, without anyone deciding to, into shared infrastructure it was never designed to be.

What it is missing

What it now lacks is everything a real system has. There is no audit trail, so you cannot see who changed what or undo a bad edit. There is no validation, so a typo becomes a fact. There is no separation between the people who need to read the numbers and the people who can break them. And there is no backstop when the one person who understands it is unavailable.

The fix is smaller than the fear

The good news is that the fix is almost always smaller than the fear. The logic inside the workbook is usually sound. It captures real knowledge about how the business runs, often better than any official document. What it needs is not replacing but rehoming: moving that logic somewhere with history, permissions, and a backup, so the knowledge survives a wrong paste or a holiday.

That move does not have to be dramatic, and it rarely means throwing the spreadsheet away on day one. It means recognising that the workbook has been doing a system's job without a system's safety net, and giving it one before the day it lets you down.

If you have a workbook like this in mind, that is a good place to look first. Our automation audit will help you see how exposed it really is, and what the smallest safe step away from it looks like.

When a spreadsheet tips over

The tells that a workbook became infrastructure.

6 tabs
90 formulas, one person understands
1 wrong paste
No version history to recover
1 person
Knows how it all connects
Not the tool
The single point of failure is
Keep reading

You might also like.

Short pieces on the patterns we see across the work, and the principles behind how we price it.

How we think

The ads are not the problem, the machine behind them is.

Turning on ads before the funnel behind them converts is the most common way businesses waste money. The funnel is the machine. Ads are just the moment you switch it on.

Read the post →
Industry observations

Most automation projects fail before anyone writes code.

When an automation project goes wrong, the tool usually gets the blame. The real cause is almost always something that happened earlier.

Read the post →
Worked examples

What we built when sales spent nine hours a week on lead routing.

A 30-person SaaS sales team was losing nine hours a week to lead routing. Here is what the fix looked like in practice, and what we deliberately did not build.

Read the post →
How we think

The list of automations is a roadmap, not a shopping basket.

Most automation pitches arrive as a menu of pick-three options. We deliver them as a sequenced plan, ordered by where the hours actually come back.

Read the post →
Patterns we see

Three signs the boring bit of your day is fixable.

A short diagnostic for people who suspect the boring layer of their day is fixable but cannot quite name what it is.

Read the post →
How we think

Why we count the hours before we touch a tool.

The most useful number in an automation project is not a price. It is the hours a task quietly eats every week.

Read the post →
Finance and accounting

Cutting the monthly close from eight days to two.

A 28-person fintech was losing the first eight working days of every month to close. Three small connections later, the close runs in two days and around 120 hours a month come back across the team.

Read the case study →
Customer service and support

Triaging 1,200 weekly tickets before an agent opens one.

A 60-person SaaS team was getting 1,200 tickets a week, most of them in five buckets. A triage layer classifies, routes, and attaches the right macro. Real bugs surface to a human inside an hour, not five.

Read the case study →
Operations and logistics

Stopping orders falling between Shopify and the warehouse.

A 45-person homeware brand was pushing orders to its 3PL via twice-daily CSV exports. A webhook-driven feed replaced the CSVs and the oversells stopped.

Read the case study →
Professional services

Client onboarding in 20 minutes, not two hours of forms.

An 18-person law firm was spending two hours of partner or paralegal time on every new matter. One form, one routing layer, and the partner signs a pre-populated letter while the rest is already done.

Read the case study →
E-commerce and DTC

Syncing 4,000 SKUs across Shopify and Amazon every hour, not every Tuesday.

A 12-person DTC pet brand ran its inventory sync as a weekly manual job. The hourly automatic sync took the manual loop away. Tuesday came back.

Read the case study →
Free tool

Not sure where to start? Take the 5-minute audit.

11 questions, your automation score, an estimated time-loss number, and three personalised recommendations. No email needed to see your results.

Take the free audit → 5 minutes · 11 questions · free
The Sprint

A focused workshop. A written report you keep.

A focused day mapping your operations and scoring the automation opportunities. The ranked plan you can act on, with us or without, follows. From £1,500.

See how a Sprint works → Focused workshop · written report · yours to keep
Talk to us

Want to chat?

Two ways in. Send us a message about what's broken, or book a 30-minute call if you already know what you'd like built. Both reach the same person.

Hours
Mon to Fri, 09:00 to 18:00 UK
Reply
Within one working day

Drop us a line.

We'll get back within one working day, usually faster.

By submitting, you agree to us getting back about this enquiry. No newsletter.