Audit Illusion And Late Certainty

How Spreadsheet Governance Starts

Month End Truth Manufacturing

Coherence Is Not Accuracy

Three Symptoms To Watch

Manual Control Debt Defined

An Ordinary Scenario Reveals Fragility

Why Spreadsheets Break At Scale

Moving Certainty Earlier By Design

Redesign The System And Wrap

SPEAKER_00 0:46

I need you to cast your mind back to the last time you opened a shared drive and were confronted with a list of files that included something along the lines of governance underscore pack underscore final dot XLSX. Governance underscore pac underscore final underscore V two dot XLSX. Governance underscore pack underscore final underscore actual dot XLSX and in a folder someone thought was helpfully labeled current governance underscore pac underscore final underscore V4 underscore use underscore this underscore one underscore March dot XLSX. If you have never experienced this, congratulations. You are either very new to the sector or operating with a level of file discipline that I would genuinely like to study. For everyone else, and I suspect that is most of us, that folder is not just a minor administrative irritation. It is a governance signal. It is telling you something specific about your organization's relationship with certainty, about how facts are created, about when truth becomes knowable, and crucially, whether that is before or after the decisions that depend on it are made. Because here is what that folder of spreadsheets is actually saying. The organization does not have a single version of reality. It has multiple working versions, each locally maintained, each plausible within its own context. And the process of determining which one is true and therefore the process of governing is a reconciliation exercise. It happens at month end or under pressure. Or when someone important asks a question that requires an answer that the system cannot immediately provide. That is the spreadsheet governance trap. And today we are going to take it apart. Welcome back to the RTO Superhero Podcast. I'm Angela Connell Richards, episode 16 of the podcast, episode 5 of the Governance Shift series. Last week we talked about the audit illusion, the self-reinforcing loop where mobilization under scrutiny substitutes for continuous governance and where the ability to perform well at audit time masks the structural lateness that lives between audits. I finished that episode with a tease that the spreadsheet is, in many organizations, the infrastructure that makes the audit allusion feel necessary. That when the operating record lives in files that require reconciliation before anyone can agree on what is true, mobilization is not a choice. It is the only option. Because certainty cannot be retrieved from the system. It has to be manufactured. Today we are going to examine how that infrastructure develops, why it feels so reasonable at every stage of its development, and why it creates a specific kind of governance, fragility that gets progressively more expensive as organizations grow. I will say upfront, I am not here to tell you to stop using spreadsheets. Spreadsheets are useful tools. They are fast, flexible, and require no implementation project. The problem is not the spreadsheet. The problem is what the spreadsheet is standing in for and what happens to governance when it does. Spreadsheet governance does not begin as a failure. It begins as a solution. And understanding how it develops is important because it explains why it is so persistent and why pointing at it and saying this is not best practice rarely produces change on its own. In the early stages of an RTO, spreadsheets are genuinely the right tool. The organization is small, the team is tight, everyone knows what the numbers mean. There is no meaningful gap between the spreadsheet and the reality it describes, because the person who built the spreadsheet is also the person doing the work it tracks. The file is live, the numbers are current. The version problem does not exist yet, because there is effectively only one version, the one on the screen of the person who runs everything. Then the organization grows. A second trainer joins, a second qualification is added, a second delivery site opens, and the spreadsheet, which was built for one person managing one program, starts to carry more than it was designed for. New columns are added. New tabs appear. Someone creates a copy so they can work on it without disrupting the live version. And then someone else does the same thing. And suddenly there are three files that are meant to represent the same reality, maintained by three different people on slightly different schedules, with slightly different understandings of what active learner means in column D. At this stage, the organization does not usually notice it has a governance problem. It notices it has a coordination challenge. And the solution to a coordination challenge is more spreadsheets, a master file, a reconciliation tab, a monthly alignment meeting where the numbers are brought into agreement before the governance pack is produced. That monthly alignment meeting is the moment the trap closes because what the organization has just designed without intending to is a system in which truth is a product of reconciliation, rather than a property of the operating record, in which governance facts are created at month end by bringing competing versions into agreement rather than retrieved from a system that holds a single current view. The timing consequence is direct. If governance facts are created at month end, they are never available before month end. Decisions that require those facts, including decisions about whether operational conditions are drifting in ways that require intervention, are deferred until the reconciliation is complete. Certainty arrives after the operating period it describes has already passed. That is what the spreadsheet governance trap actually costs, not efficiency. Time. Spreadsheet governance can feel most in control in the very organizations where control is actually weakest. Let me explain why. When an organization runs on spreadsheets, it becomes very skilled at producing coherence. The monthly reconciliation exercise over time becomes a well-practiced discipline. People know their role in it. They know which columns to update, which definitions to apply, which numbers need to be brought into alignment before the pack goes to the board. The organization becomes fluent in reconciliation. It can produce a clean, coherent, internally consistent governance pack on a reliable schedule. And that fluency feels like capability. It looks from the outside and often from the inside like a well governed organization. The numbers are tidy, the narrative is coherent, the pack arrives on time. Governing persons receive a consistent account of the organization's condition. What they do not receive and what the reconciliation process has, in a sense, removed is the variance. The small differences between what different parts of the organization were seeing in real time before they were smoothed into agreement. The early signals that got averaged away, the definition drift that got papered over, the edge conditions that got rounded to fit the template. The pack is coherent because the raw material that would have disturbed its coherence has been filtered out in the process of producing it. And so governing persons are confident, reasonably confident, based on the information they have received, in a picture that may no longer reflect the actual condition of the organization with any precision. Coherence is not the same as accuracy. A reconciled governance pack can be perfectly coherent and structurally misleading at the same time, because reconciliation removes the variance governance needs in order to act early while producing the consistency that makes governance feel settled. Part three. The three symptoms. I want to name them clearly because if you can see them in your organization, you have a precise diagnosis of where your governance timing problem is being produced. The first symptom is version proliferation. This is the folder I described in the cold open. Governance underscore pack underscore final underscore V4. Underscore use underscore this underscore one. When versions multiply, it is not primarily an administrative problem. It is a signal that the organization has no single authoritative source of truth. And that governance facts require negotiation before they can be established. The practical consequence? Teams spend increasing proportions of their time determining which numbers are correct rather than acting on them. Comparability erodes. The effort required to establish a shared baseline before any decision can be made increases with every version added to the system. The second symptom is definition drift. This one is subtler and more consequential than version proliferation. Core terms completion, active learner, outstanding assessment, validated tool, industry engagement begin to carry different meanings across different files and functions. Each definition is internally consistent. None of them are the same. And when the governance pack tries to aggregate across these different definitions, the result is not an accurate picture of the organization. It is a composite of multiple different pictures that happen to use the same column headers. I have sat in governance meetings where three different functions were asked what percentage of learners in a particular program were on track and received three answers that differed by 15 percentage points. Nobody was wrong. They were just measuring different things with the same word. That is definition drift. And under scrutiny, when a regulator or funder asks a question that requires a consistent answer, definition drift does not stay inside the governance pack. It becomes an evidentiary problem. The third symptom is fragmentation of the evidence chain. This is the most governance critical of the three. When decisions are made in meetings or messages and supporting artifacts sit across multiple drives and trackers maintained by different people with different update schedules. The sequence that connects signal, decision, and outcome does not exist in a retrievable form. It can be reconstructed. Given time and effort and access to the right inboxes and file versions, a plausible account can be assembled. But the contemporaneous record, the proof that governance acted while control was possible in real time, as conditions were forming, is not there. And under continuous assurance, reconstruction is not the same as retrieval. The question is not whether the organization can assemble the story now. It is whether the evidence existed then. Part four. Manual control debt. I want to introduce a concept from the book that I think gives this dynamic a useful frame. Manual control debt. Manual control debt is the accumulated liability created when an organization relies on workarounds, spreadsheets, inbox approvals, reconciled packs, and individual knowledge to simulate control. Each workaround is a small, local solution to an immediate problem. Someone needs to track trainer credentials. A spreadsheet is created. Someone needs to monitor assessment turnaround, another spreadsheet. Someone needs to reconcile the enrollment data between the SMS and the finance system. A monthly manual process is established. Each of these solutions works. In the short term, in isolation for the problem it was designed to solve, it works. The debt accumulates not in any single solution, but in their combination, in the multiplying versions, the diverging definitions, the growing reconciliation burden, and the increasing distance between when something happens in operations and when governance can be certain about it. Manual control debt is tolerable in calm periods. In fact, it is often invisible in calm periods because the organization has developed the coordination practices needed to manage it. The monthly reconciliation runs smoothly. The master spreadsheet is maintained. The pack goes out on time. Everything appears to be working. Under pressure, the debt charges interest. Exceptions multiply. Time compresses. The reconciliation that worked fine when conditions were stable becomes unreliable when conditions are moving quickly. Because the gap between what is actually happening and what the spreadsheet reflects widens faster than the update schedule can close it. Certainty moves further from the moment decisions are being made, and governance becomes progressively less able to act in time. Manual control debt is paid in time, and time in a continuous assurance environment is exactly what the organization cannot afford to spend reconciling what it already should have known. The reason this concept matters is that it reframes the spreadsheet problem from a technology question to a governance timing question. The issue is not that spreadsheets are inefficient, it is that they push certainty later. And late certainty is the defining characteristic of fragile governance. Let me give you the scenario. This one is deliberately ordinary, because the trap is most consequential in ordinary conditions, not crisis ones. A composite RTO has grown steadily over several years. It now operates across multiple qualifications and two delivery sites. Governance reporting is produced monthly from a set of well-maintained spreadsheets. A trainer matrix, an assessment tracker, a validation register, and a financial worksheet projecting revenue by intake. Each month a small team spends a day and a half reconciling these sources into a governance pack. The process is experienced. The pack is clean. Leadership has confidence in it. Nobody would describe this as a governance problem. From the inside, it looks like a well-run organization, applying appropriate rigor to its governance reporting. Over a new intake cycle, however, a gap begins to form. The trainer matrix and the assessment tracker are updated on different schedules. A learner appears under slightly different identifiers in the enrolment system and the assessment tracker, a minor data entry inconsistency, the kind that gets corrected at month end. The validation register uses a different definition of completed than the assessment tracker. Each difference is small. Each is resolved at reconciliation time. The pack remains clean. Then mid-cycle, a contract manager asks a question. Not an unusual question. Just can you show me the current status of learners in cohort three using the same definitions across enrollment, delivery, quality, and finance? The organization cannot answer immediately. Not because the data does not exist, but because the data exists in four places, using four slightly different definitions, updated at four different points in time. Answering the question requires a manual alignment exercise. Columns are reconciled, definitions are clarified. Someone calls someone else to confirm which version of the assessment tracker is current. An answer is eventually produced, accurate, defensible, and constructed entirely in response to the question, rather than available before it was asked. The realization in that moment is not that the data was wrong. The realization is that truth in this organization, with this operating infrastructure, cannot be retrieved. It must be assembled. And if truth must be assembled, then governance facts do not exist at the point decisions are being made. They exist at the point someone asks for them, which is definitionally too late for governance to have acted on them. The organization is not failing. It is functioning exactly as it was designed to function. The design produces late certainty, and late certainty is the infrastructure of the audit illusion. Because when certainty is never available before month end, mobilization is the only mechanism through which it can be produced under pressure. Part 6. Why Spreadsheets Break at Scale. I want to be specific about why this problem intensifies as organizations grow. Because the experience of scale often catches people off guard. The spreadsheet approach that was genuinely adequate at one size becomes structurally insufficient at a larger size. Not gradually, but with a kind of threshold effect where the system that was working suddenly stops working, and the reason is not immediately obvious. The first pressure is volume. More qualifications, more cohorts, more sites, more staff, more partners. Each addition extends the spreadsheet or creates a new one. The reconciliation burden grows. The time between when something happens and when it is reflected in the governance pack extends. The monthly truth manufacturing ceremony, which is what a reconciliation exercise effectively is, takes longer and requires more people and still produces a picture that is already somewhat out of date by the time it arrives. The second pressure is ownership dilution. In a small organization, the person who built the spreadsheet is also the person who interprets it and who notices when something looks wrong. At scale, the builder, the maintainer, and the interpreter are often different people. And the tacit knowledge that makes the system legible does not transfer cleanly. Columns that mean something specific to their creator mean something slightly different to the person updating them two years later. Definition drift accelerates. Version control becomes a full Full-time job that nobody has been assigned to do. The third pressure is the key person dependency. When the operating record lives in spreadsheets that are maintained by individuals, the legibility of the governance system is a function of those individuals' presence. A resignation, a leave of absence, an illness, and the organization discovers that its governance infrastructure is not a system. It is a person. And persons, unlike systems, are not reliably available when governance needs them most. I have seen this play out in audit context more than once. A regulator asks a technical question about the evidence trail for a specific cohort. The person who maintained the relevant spreadsheet left six months ago. The file exists, the knowledge required to interpret it does not. The organization can reconstruct over several days an answer that is probably correct. What it cannot do is demonstrate that governance was operating in real time while the cohort was live. Scale does not create the spreadsheet governance problem. It removes the informal buffers, proximity, shared context, key person knowledge that made the problem tolerable at smaller size. Part 7. What to do about it. Because the answer is not, as I sometimes hear proposed, implement a better spreadsheet system. And it is also not rip everything out and replace it with enterprise software. Neither of those responses addresses the actual problem. The actual problem is that certainty arrives later than decisions require. Everything that follows from that is a symptom. The solution is to move certainty earlier so that the operating record reflects current conditions in real time and governance facts do not require manufacturing. The first step is the most unglamorous, but it is also the most important. Stabilize your definitions. Pick one meaning for active learner. Pick one meaning for completed. Pick one meaning for on track. Write it down. Apply it consistently across every system that touches those concepts. This sounds trivial, it is not. Definition drift is the root of the version proliferation problem, because when the same term means different things in different places, reconciliation is not a coordination failure. It is a necessary translation exercise. Removing the drift removes the need for translation. The second step is to identify what single version of truth your governance actually needs and make that version continuously available rather than periodically manufactured. This does not necessarily require a major system change. It requires being specific about which facts governance needs to make timely decisions and designing the operating record so those facts are accessible without a reconciliation exercise. In many organizations this is achievable with existing tools if the intention is there. The third step is to treat reconciliation as a warning sign rather than a governance mechanism. Every time your team spends time bringing numbers into agreement before governance can use them, that is the sound of manual control debt being serviced. Track where it happens, how long it takes, and what decisions were deferred while it was occurring. That tracking will show you exactly where your governance timing is being produced by your infrastructure and exactly where the design work needs to go. The book maps this across all eight critical drivers, showing where each driver's governance visibility is most commonly undermined by spreadsheet infrastructure and what a more integrated operating record would need to hold. It is the most practically detailed section in the book, and I think it is where the most immediate value sits for leaders who are ready to move from diagnosis to design. I want to close with the observation that I think reframes this whole conversation. The spreadsheet is not the problem. The spreadsheet is a symptom of a design choice, a choice usually made implicitly and incrementally rather than deliberately, to allow governance certainty to live in reconciliation rather than in the operating record. That choice is understandable. It happens in almost every organization that grows beyond its original size without pausing to ask whether its governance infrastructure has grown with it. But the choice has consequences. It pushes truth later. It makes the signal chain fragile. It makes the governance visibility gap wider. It makes the audit illusion necessary. It accumulates manual control debt that charges interest under pressure. And it produces organizations that can describe themselves clearly and struggle to govern themselves in time. The good news, and I mean this genuinely, is that it is a design choice, which means it can be redesigned. Not all at once, not in a single project, but deliberately, specifically, at the points in your operating model where reconciliation is currently doing the work that contemporaneous certainty should be doing. That is what the next phase of this series is about. Next week, we are going to talk about the governance divide, the emerging split between providers who govern early through design visibility and those who integrate only when events force it. We have been building toward this all series. The spreadsheet governance trap is one of the clearest structural explanations for why organizations end up on the fragile side of that divide. Not through bad intent, but through infrastructure that was never designed to support the governance standard the environment now demands. The governance shift in vocational education, out June 2026. The chapter on the spreadsheet governance trap is chapter 8, and the manual control debt analysis that follows it in chapter 9 is where the most useful diagnostics sit. Links and the free governance scorecard are in the show notes. The phrase this is the master spreadsheet rarely holds for long because the system it represents is already distributed. Someone, somewhere, has already made a copy. The question is what that copy is costing your governance, and whether it's a cost you've been too busy reconciling to notice. You have been listening to the RTO superhero podcast. I'm Angela Connell Richards, go be governable.