Your business runs on folklore. That's the problem.
If your business would fall apart when one spreadsheet goes missing, you don’t have a scrappy operation.
You have a fragile one.
And most scale-up SMEs don’t notice they’re fragile until something small - something that shouldn’t be a problem - exposes the whole thing.
Someone goes on holiday. A board paper needs numbers by Friday. A new hire asks where to find revenue by client. A customer wants an honest update and three people give three answers.
That’s the moment the truth lands: the business is held together by a stack of workarounds, and by the people who happen to remember how it all fits.
The stack everyone has and nobody documents
The real tech stack of a £3m agency or scale-up SME usually looks like this:
Google Sheets. HubSpot. Xero. Zapier. Monday.com. Maybe Slack doing more work than Slack should.
All fine tools. The tools aren’t the problem.
The problem is what grows in the gaps between them:
- Spreadsheets doing the job of a database.
- Two CRMs, because you migrated to the new one but never turned the old one off.
- Reports built by exporting CSVs and stitching them together by hand.
- Numbers that are “close enough” - until a board meeting, an acquisition conversation, or a client dispute.
This is the default state of a growing business. Almost nobody escapes it accidentally.
It’s also the reason so many founders feel that growth means more checking, more chasing, more clarifying, and more time spent answering the same five questions about what’s actually true.
That feeling has a name. It’s infrastructure debt. And like all debt, you pay interest on it whether you acknowledge it or not.
What infrastructure debt actually is
Infrastructure debt is the accumulation of operational shortcuts that kept you moving when you were small - and now slow you down.
It’s the gap between how a business of your size should run and the improvised plumbing keeping it upright.
You won’t see it on the balance sheet. You’ll see it as:
- “Reporting takes ages.”
- “The numbers don’t match.”
- “Only Priya knows how that works.”
- “Don’t touch that Zap, I don’t know what’ll happen.”
You can absolutely get to seven figures on improvised plumbing. Plenty of businesses do.
But the moment you try to scale past it, the interest payments start showing up everywhere.
The most obvious symptom: reporting nobody trusts
Here’s a scene I see almost every month.
A founder asks for a basic view of the business. Pipeline, delivery status, capacity, revenue. Nothing exotic.
What follows is a ritual.
Someone exports from HubSpot. Someone pulls invoices from Xero. Someone takes a screenshot of Monday. Then a senior person - usually one of the most expensive people in the building - spends two hours stitching it together in a spreadsheet.
The result lands in your inbox. And in the back of everyone’s mind, including yours, is the same quiet thought:
This is probably not very accurate.
That’s an ops problem that only shows up at the reporting layer.
Because once leaders stop trusting the numbers, they stop using them. They go back to gut feel and Slack threads and whoever spoke loudest in the last meeting. And that’s where chaos starts to feel normal.
The scarier symptom: one person is the system
A few years ago I worked with a business where one person knew how everything ran.
Not officially. Not on the org chart. Just - in practice. She was the person you went to when the numbers didn’t add up, when a process broke, when something needed pulling together for a client by end of day.
Then she went on holiday. A real one. Two weeks.
The business lost a week. Genuinely lost it. A week of “where is it?”, “who has it?”, “which version is current?”, “why doesn’t this match what we sent the client?”
When she came back, half the questions resolved in a single sentence:
“Oh, I’ve got a spreadsheet that does that.”
A spreadsheet only she used. Built for her own sanity. Never shared, never documented, never replicated. Everyone else who’d known it existed had already left the company.
That’s infrastructure debt in its purest form. A hidden tool, running the business from the shadows.
And let me be blunt: if your business loses a week because one person is at the beach, you don’t have a scalable company. You have a dependency wearing a job title.
The Folklore Test
If your operations rely on memory, you don’t have operations. You have folklore. (no, not the Taylor Swift album.)
Pick any process that matters - pipeline, billing, delivery handover, resourcing, anything - and ask your team four questions:
- Where does this data live?
- Who owns it?
- What happens if it breaks?
- Could someone hired next month run it without supervision?
If the answers are “it depends”, “I’ll check”, “only Priya knows”, or nervous laughter, you’ve found a debt position.
You don’t need a consultant to tell you what to do next. You just need to stop pretending it’s fine.
What infrastructure debt is actually costing you
Even when revenue is strong, you’re paying for this. Just in places that don’t show up clearly.
Founder dependency
You become the escalation path for everything. The translator between systems. The only person who can confidently say what’s true. Not because you want to be - because nothing else can stand on its own.
This is the cost most founders feel daily and never quite name. It’s the reason your inbox doesn’t get smaller as the business grows.
Delivery drag
Work doesn’t fail. It just sticks. Handoffs take longer than they should. Approvals stall waiting for someone to confirm something that should be visible. People are busy, but progress is heavy.
Drag is the most expensive thing in a service business and the hardest to see, because nothing is ever obviously broken.
Hiring friction
New hires take six months to be useful instead of six weeks. Not because they’re slow - because “how we do things” lives in three people’s heads and four private spreadsheets.
Every hire pays the onboarding tax. You pay it twice when they leave.
Margin erosion
This is the one founders consistently underestimate.
Manual reporting eats senior time. Reconciliation meetings eat everyone’s time. Rework eats delivery margin. Mistakes from bad data eat client trust, which eats retention, which eats LTV.
None of it shows up as a line item. All of it shows up in the P&L.
Why this happens - and why it’s not your fault yet
This isn’t incompetence. It’s the natural physics of growth.
When you were five people, everyone held the business in their head. Processes were verbal. Tools were flexible. Speed mattered more than structure, and that was the right call.
Then you grew. More clients, more delivery, more hires, more handoffs, more moving parts.
So you patched. A spreadsheet here. A Zap there. A second CRM because the first one got messy and switching felt cheaper than fixing. A “temporary” report that’s now in its fourth year.
Each individual patch made sense at the time. The problem is the sum of them. Eventually nobody can describe what the system is. They just know the work somehow gets done.
It stops being your fault when you didn’t know. It starts being your problem the moment you do.
The trap: buying a tool instead of making a decision
When the cracks start showing, the reflex is almost always:
“We need a better tool.”
Sometimes you do. Usually you don’t - at least not first.
Most scale-ups don’t have a tooling problem. They have a clarity problem that no tool will solve on its own. A new platform won’t fix:
- Unclear ownership.
- Inconsistent definitions of basic things (“what counts as a qualified lead?”).
- Multiple sources of truth that all feel slightly correct.
- Data hygiene that depends entirely on the goodwill of whoever entered it.
- Processes that only exist in Slack threads.
A shiny new platform laid on top of all that just gives the chaos a nicer interface. And now you’ve paid for it.
The work has to start earlier.
How to fix it without a dramatic rebuild
This is the unglamorous part. It’s also the part that works. Do these in order - the order matters more than people think.
1. Pick one winner per category
Not one tool for everything. One winner for each kind of truth: pipeline, delivery status, capacity, revenue, reporting.
If two systems might be correct, neither of them is. Pick the winner, write it down, and turn the other one off - or at least demote it so nobody confuses it for authoritative.
This is the step most teams skip because it requires saying no to something someone built. Do it anyway.
2. Make reporting a byproduct, not a project
If your monthly numbers require manual copy-paste, you don’t have reporting. You have arts and crafts.
The bar is this: reporting comes from agreed sources, with consistent definitions, without heroic effort.
That usually means simplifying what you measure (most scale-ups track too many things badly instead of a few things well), standardising how data is entered, and killing the duplicate systems you’ve been politely tolerating.
3. Treat your automations like production systems
Because they are.
If a Zap is doing anything that would matter if it broke, it needs an owner, basic documentation, and some form of monitoring - even if “monitoring” is just a weekly check. Mystery automations are how businesses drift into silent failure: the Zap broke in March, nobody noticed until June, and now you’re explaining to a client why their renewal email never went out.
4. Kill the private spreadsheets
This one is touchy, because private spreadsheets are usually built by your most competent people trying to survive the chaos you didn’t fix.
The rule still holds: if it’s business-critical, it can’t be private. Either promote it into the system with an owner and documentation, or remove the need for it entirely.
If you can’t do either, you’ve just confirmed it’s load-bearing - and load-bearing things shouldn’t live on one laptop.
5. Assign ownership by name
Not “ops owns it”. Not “the team owns it”. One named human per system or process, who defines the rules, keeps the data clean, approves changes, and is responsible when entropy creeps in.
This is how you stop the slide back into folklore. Without it, every fix you make has a half-life of about six months.
What “good” actually looks like
Good ops in a scale-up SME is much simpler than perfect tooling.
- Everyone knows where the truth lives.
- Reporting doesn’t need reconciliation rituals.
- Systems don’t collapse when one person takes a fortnight off.
- Automations have owners and documentation.
- Delivery runs without heroics.
Calm systems create calm leaders. Chaos creates indispensable founders. And “indispensable” is a description of a business that can’t grow past you.
If you want to run the Folklore Test against your own business, there’s a printable scorecard you can use. Same four questions, applied across your critical processes, with a simple read on where the debt is sitting.
