The compound cost of operational debt

Most founders can describe the feeling before they can name the cause. The team is larger but decisions are slower. Systems exist but nobody fully trusts the data. Processes work, in the sense that things get done — but they depend on specific people knowing which workaround to use, which spreadsheet to check, and which exception to approve. A 2023 Jitterbit study — surveying 400 founders and C-suite executives at venture-backed companies across Europe — found that 87% cite manual data processes and data silos as the primary barriers to growth. Not competition. Not capital. The internal machinery of the business itself.

What these leaders are describing has a name, though few use it: operational debt.

What operational debt actually is

The concept borrows from software engineering, where technical debt describes the accumulated cost of shortcuts taken to ship code faster. Operational debt works the same way, but across the entire business. It is the accumulated cost of every workaround that replaced a documented process, every tool adopted without integration, every decision that routes through a person instead of a workflow, and every piece of institutional knowledge that lives in someone's head rather than in a system.

Steve Blank introduced the related concept of "organisational debt" — the accumulation of deferred decisions about structure, roles, and culture. Operational debt is broader. It encompasses process design, system architecture, data quality, decision logic, and the readiness of the operating model to support what the business needs to do next. Where technical debt sits in the codebase, operational debt sits in the operating model — and unlike technical debt, it rarely has anyone assigned to measure it.

The distinction matters because operational debt is not a metaphor. It is a quantifiable cost that accumulates in every scaling business as a natural consequence of growth outpacing operational design. The tools were chosen for speed, not integration. The processes were improvised because there was not time to design them. The decisions run through the founder because nobody else has the context. Each of these was the right call at the time. The debt is not in the individual decision — it is in the accumulation.

How it accumulates

Operational debt grows fastest at inflection points: when headcount doubles, when new channels launch, when the founder steps back from day-to-day operations, when the business enters a new market. Each inflection point stretches existing processes — and if the response is a workaround rather than a redesign, the debt deepens.

The evidence on workaround proliferation is striking. ProcessMaker research estimates that 40% of an average team's time is consumed by shadow tasks — activities that are undocumented, untracked, and invisible to management. These are not minor inefficiencies. They are the actual operating model: the way work really happens, as opposed to the way anyone thinks it happens. A JumpCloud and Gartner analysis found that 42% of company applications are shadow IT — systems adopted outside formal governance, creating data silos that no one fully maps.

The pattern compounds. A tool adopted to solve one problem creates a data silo. The silo requires a manual workaround to bridge it to another system. The workaround depends on a specific person who knows how to do it. When that person is unavailable, the process breaks. When the team grows, the workaround scales poorly. When the business tries to automate, it discovers that the process it wants to automate has never been documented — because the workaround was never supposed to be permanent.

NextProcess research found that 94% of business spreadsheets contain errors. In organisations where spreadsheets serve as the connective tissue between systems — which, in a business of 30 people growing at pace, is most of them — those errors compound through every process that touches them.

The cost nobody measures

The financial weight of operational debt is substantial, even when no one is tracking it. A Parseur 2025 survey found that manual data entry alone costs US companies $28,500 per employee annually — more than nine hours per week spent transferring information between systems by hand. For a 50-person business, that represents over $1.4 million in annual cost before anyone has measured the knock-on effects: the decisions delayed by unreliable data, the errors introduced by manual transfer, the time spent resolving inconsistencies downstream.

The broader picture is harder to calculate and harder to ignore. Research on the Cost of Poor Quality — a concept from the Six Sigma and Juran tradition — estimates that poorly designed processes consume 15–20% of revenue in the average business, through rework, error correction, and wasted effort. For a business turning over £5 million, that is £750,000 to £1 million a year absorbed by operational friction before anyone has named it as a problem.

And yet, Geckoboard research found that 49% of SME owners have failed to identify any key performance indicators at all. The businesses carrying the most operational debt are, by definition, the ones least equipped to measure it. The cost is real — it shows up in slower execution, frustrated teams, and decisions made on incomplete information — but it sits below the surface of the metrics that leadership watches.

McKinsey's Organisational Health Index, based on more than eight million survey responses across 2,600 organisations, found that companies in the top quartile for organisational health deliver three times greater total shareholder returns than those in the bottom quartile. Organisational health — the systems, processes, and operating disciplines that sit beneath financial performance — is not a secondary concern. It is a leading indicator of value creation.

What it does to growth

The relationship between operational debt and growth failure is direct. The Startup Genome Project found that 74% of high-growth startups fail due to premature scaling — not from lack of demand or capital, but from operations that cannot support the scale the business has reached. The pattern is specific: businesses that scale headcount and revenue before their operational foundations can absorb the increased complexity hit a ceiling that no amount of hiring or investment can move.

Every new customer, order, market, or hire multiplies the cost of existing operational debt. A process that generates three exceptions per week at 20 people generates 30 exceptions per week at 60. A data inconsistency that one person can manually correct becomes a systemic reliability problem when the team relying on that data triples in size. The business does not fail because it stops growing. It fails because the operational cost of growth outpaces the commercial value it creates.

This is why founder-dependent businesses attract lower valuations from strategic acquirers — not because the business is performing poorly, but because the performance depends on architecture that has not been designed to sustain itself. The pattern of constraint in scaling businesses is well documented: the same bottleneck shows up in the founder's calendar, the operating model, and the expansion strategy. Operational debt is the mechanism through which that constraint tightens.

Why it matters now more than it ever has

In a pre-AI world, operational debt was a drag on efficiency — costly but survivable. In 2026, it is the primary reason AI investments fail. AI amplifies what exists. If what exists is fragmented, undocumented, and disconnected, AI makes that worse, faster.

MIT research found that 95% of generative AI pilots fail to create measurable business value. RAND Corporation analysis of the $684 billion invested globally in AI in 2025 found that more than 80% — over $547 billion — failed to deliver its intended value. These are not technology failures. They are operational debt surfacing at the worst possible moment: 84% of AI project failures are attributed to leadership and organisational issues, not technology limitations.

The logic is straightforward. AI tools need documented processes to follow, clean data to learn from, integrated systems to access, and clear decision logic to apply. Operational debt is the absence of all four. Every business deploying AI onto a foundation of workarounds, manual data bridges, and undocumented tribal knowledge is automating the problem, not solving it.

This is not a future concern. Gartner predicts that 60% of AI projects will be abandoned by the end of 2026 due to inadequate data readiness. The businesses that address their operational foundations before investing in AI will capture the value. Those that invest in AI before addressing operational debt will pay for the debt twice — once in the accumulation, and once in the failure.

A lens, not a verdict

Operational debt is not a failure. It is an inevitable consequence of building under uncertainty. Every business that has grown past its initial operating model carries some. The decisions that created it — choosing speed over documentation, workarounds over redesign, founder judgment over formalised process — were almost always the right calls at the time.

The opportunity is in seeing it clearly. In naming it precisely enough that it stops being a vague sense that things are harder than they should be, and starts being a measurable feature of the operating model that can be addressed deliberately. The businesses that treat operational debt with the same rigour they apply to financial debt — understanding where it sits, what it costs, and what it prevents — build the foundation for everything that comes next: sustainable growth, effective AI adoption, successful expansion, and an operating model that works without the founder in every room.

The first step is the simplest: name it. The second is to measure it. Everything after that is architecture.