Operational house of cards: when the workarounds stop working

There is probably a spreadsheet somewhere in your business doing something important. There could well also be an app being built right now with AI, someone’s self-taught trial of vibe-coding.

Not a specific one, necessarily. Something that reconciles two systems that do not talk to each other, or tracks something no other tool captures, or contains complex formulas or a script that one person built over time and that nobody else has fully understood. It works. It has always worked. The fact that it exists is not a sign that anyone made a bad decision.

It is still a house of cards. And the house falls when that kind of infrastructure — the accumulated workarounds and manual connections of a business that has been moving fast — becomes a liability rather than a solution. It does not usually announce itself with warning signs. It announces itself with an operational failure at the worst possible moment.

How the stack gets built

Nobody builds an operational house of cards on purpose. It accumulates through a series of individually sensible decisions made under time or cost pressure.

You needed to track something and there was no system for it, so you built a spreadsheet. You needed two tools to share data and the integration cost felt disproportionate, so someone set up a manual export. You needed a plan linking delivery to financial data, a supplier pricing matrix, an inventory tracker. You outgrew your original CRM but the migration felt too disruptive to run during a busy period, so you kept both running and added a reconciliation step. You hired someone who was good with these things and interested in AI, and they became the keeper of several processes that were never formally documented. Sometimes it grows arms and legs, and an application gets built that no one else knows how to run.

Each of these decisions was reasonable. That matters. Fast-growing businesses survive by solving problems quickly with the limited resources available. The founders who get things done are the ones who accept imperfect solutions and keep moving. The house of cards is the price of that innovative pragmatism, and for a long time it is a price worth paying.

At 10% growth, you start to notice the compounding.

Where founders tend to get this wrong

The most common mistake is treating this as a future problem — sorting the systems out after the fundraise, after the next quarter, when there is more capacity. The logic is understandable. The practical effect is that the future never quite arrives, and the problem grows while it waits.

The second mistake is reaching for a new system before understanding the underlying process. Buying better software and implementing it on top of a broken process produces a more expensive version of the original problem, replicating all the underlying issues. Systems do not fix processes. Processes need to be right before systems can support them well.

The third mistake is underestimating the people dimension. Every system change affects the people who use it. A finance team that has worked in a particular platform for three years will find the adjustment difficult. An operations team whose workflows are built around manual steps will, consciously or otherwise, find ways to reintroduce those steps even after a new system is in place. Implementation without change management tends to return more or less to its previous state.

Start with one question, asked of every critical process

Before any of this becomes a systems project, it is worth mapping the critical processes in the business and asking one thing of each: what happens if the person who runs this leaves tomorrow?

Is there documentation? Is there a system? Or is there a spreadsheet on someone’s desktop that only they fully understand?

It is a deliberately unglamorous exercise, and it does two useful things at once. It gives you an honest map of where the concentration risk actually sits, which is rarely where people assume. And it separates the processes that merely feel fragile from the ones that would genuinely stop the business, which is the difference between a long list of concerns and a short list of priorities.

Do it before you look at software. The answer will change what software you need.

What comes next

The map is the start. The full article covers how to identify and cost the manual steps honestly — not just the time they take today, but the time they will take when the business is twice the size; how to tell a system problem from a process problem, because conflating them is an expensive way to discover the distinction; what scalable operations actually look like; and why operational and technology advisory is an area where generalist advisors often reach the edge of their competence without flagging it.

At Link Stone Advisory our advisors include senior executives with hands-on operational experience in scaling businesses, working alongside tech leaders. We help founders identify where technical debt sits, what it is costing, and how to address it without disrupting the business that is already running. We do not build code, so the advice you get is objective and independent.

We offer a free initial consultation in the strictest confidence. If you want to expect more from your board and professional advisors, contact us.