Why Your Best Process Documentation Is Already Obsolete
Process documentation gets written at the moment of greatest accuracy. Someone maps the steps, checks them with the people who do the work, and collects the sign-off, often right after a launch or a clean review. For one brief stretch, the page and the practice match.
Then the work moves on. A system gets a new field. A customer needs an exception. A team member finds a faster route and passes it along over lunch. None of it feels big enough to record, so none of it is. The document stays exactly as it was approved, and the real process drifts away from it a little at a time.
Nobody notices, because the work still gets done. The gap shows only when something forces a comparison. An audit forces it. An acquisition forces it. So does the day a long-serving employee gives notice and takes the real process out the door.
The Day You Finish Is the Day It Starts Aging
Researchers who study organizational routines explain why. In Reconceptualizing Organizational Routines as a Source of Flexibility and Change, published in Administrative Science Quarterly in 2003, Martha Feldman and Brian Pentland split every routine into two parts. The ostensive part is the idea of the routine, the shared structure people point to when they describe how work gets done. The performative part is what specific people actually do, at specific times and places, to bring the routine to life.
The two parts keep shaping each other. What people do changes what people understand the routine to be. A written procedure captures that understanding on a single day. The doing continues, and the page does not. The paper does not address written procedures as such, so applying it to documentation is my reading.
Some of that drift is improvement. People find a better sequence, drop a step that added nothing, or work around a bottleneck, and adaptation is how organizations learn. The failure is quieter. A better way of working that never reaches the page stays private to the people who found it. It can't be taught to new hires, tested against risk, or copied by the next team. The target is not a process that never changes. It's a record that keeps pace with the change.
Why the Drift Stays Quiet
Drift produces no failure. Orders ship. Invoices go out. Customers get served. The output looks the same, so nobody has a reason to open the document and check it.
Researchers in process mining, who compare recorded process models against event data from real operations, say it plainly. Wil van der Aalst, Josep Carmona, Thomas Chatain, and Boudewijn van Dongen write in A Tour in Process Mining: From Practice to Algorithmic Challenges that process models, whether built by hand or generated from data, "may deviate from reality." That sentence carries weight. It comes from people whose field includes measuring the gap.
Three events end the silence, and each makes the gap expensive in its own way.
An audit tests what you say you do against what you do. The reviewer might be an external auditor in Europe, a regulator in Asia, or a customer's quality team in North America. The question is similar in each case, and a sound process described by an outdated document can still produce a finding. Management must then work out whether each deviation was an authorized improvement, an ineffective control, or a lapse in accountability.
An acquisition puts your documents in front of people who have no reason to be generous. A buyer's diligence team reads them to judge how the business really runs, and one mismatch tends to spread the questions to every other place. How much of the performance depends on undocumented workarounds? Can the buyer bring systems together without disrupting the work? These uncertainties do not automatically lower a valuation. Their weight depends on the transaction and on how critical the affected processes are.
A departure removes the one person who knew which steps on the page were ceremonial. The workaround, the unwritten sequence, and the right call when the system times out all leave together. The departure did not create the weakness. It exposed a dependency that management had not addressed.
When Drift Starts to Matter
Hard data on how often this happens is scarce in the sources I reviewed, but one recent study is instructive. In Workarounds as a Cause of Mismatches in Business Processes—Insights from a Multiple Case Study, published in Business & Information Systems Engineering in 2025, Christian Bartelheimer, Bernd Löhr, Malte Reineke, Agnes Assbrock, and Daniel Beverungen examined four organizations, among them manufacturers and a public university. They logged more than 110 candidate workarounds, confirmed 44, and studied 13 in depth. Staff skipped a required request for quotation, ordered from vendors by phone instead of through the company system, and placed orders without an offer. In some cases the system records still matched the documented process while the work did not.
The study covers a small set of cases and does not estimate how common workarounds are elsewhere. It does show how ordinary the drift can be, and drift takes forms that carry different weights. Some of it is cosmetic. A document names software the team retired, or routes approvals to a role that a reorganization renamed. That costs time. Two other forms are more serious. Exceptions can harden into the main path, so the procedure describes only the minority case. And controls can disappear. A check that felt redundant gets skipped once under deadline pressure, then skipped every time after, and nobody updates the page because nobody wants to announce it. A missing control is what reviewers look for, and it is the item most likely to hide between the page and the work.
The Better the Documentation, the Bigger the Risk
Most teams get this part backward. Well-built documentation is detailed, polished, and formally approved. Those same qualities make it hard to change. Updating it means a revision cycle, a sign-off, and a new version number, so people put it off. Meanwhile it looks authoritative, so people trust it. New hires follow it and get misled. Veterans spend hours answering questions the manual should answer. Leaders point to it when someone asks how things work, and the organization ends up running two versions of the same process without realizing it.
Poorly written documentation gets ignored, which limits the damage. Excellent documentation gets believed, which multiplies it. This is my judgment, not a research finding, and it is worth testing against your own files.
A Reviewed Document Is Not a Verified Process
Many organizations manage accuracy through an annual review. An owner rereads a procedure, finds the language reasonable, and approves it for another year without watching the work. The document has been reviewed. Its accuracy has not been established.
Two widely used frameworks point the same way. The Committee of Sponsoring Organizations (COSO) of the Treadway Commission describes five components in its Internal Control—Integrated Framework, first issued in 1992 and updated in 2013. Control activities are the actions set through policies and procedures. Information and communication carry what people need to do their part. Monitoring activities check whether controls are present and functioning. The framework also says internal control is about people and the actions they take, not merely policy and procedure manuals. COSO describes the original framework as widely used around the world, and whether an independent auditor examines internal control varies by jurisdiction. A written procedure supports control. It does not prove the control runs.
The International Organization for Standardization (ISO) lists documented information and the monitoring, measurement, analysis, and evaluation of performance among the requirements of ISO 9001:2026 Quality management systems — Requirements, published on September 16, 2026. ISO has withdrawn the 2015 edition and says certified organizations must move to the new one within the timeframe set by their certification cycle. For those organizations, the transition is a built-in reason to check the documents against the work.
Treat Process Documentation as a Perishable Asset
A document with no expiry date is a promise nobody is checking. Think of it as inventory with a shelf life, and manage it that way.
Name an owner who performs the process, not the person who wrote the document, and make that owner responsible for operational accuracy. Writers move on. Operators feel the drift first.
Set the review rhythm by risk and rate of change. Anything tied to software, regulation, or customer-facing steps deserves a quarterly look. Stable back-office routines can wait for an annual one.
Run a walk-through test. Ask someone who has never done the process to follow the document step by step while the person who does the work watches. Each time the operator says a step isn't done that way anymore, you have found a drift point. Count them. That number tells you how far the page has moved.
Close the loop at the moment of change. No process change is finished until the document reflects it, so make the edit path as light as the change itself. Investigate recurring workarounds before deciding whether to adopt, redesign, or prohibit them.
Technology can make documentation easier to produce and share. It can also make an inaccurate procedure look more authoritative. Validate the process before you speed up the paperwork. Where your systems record event data, compare it with the documented flow, as the process mining researchers do. Then start small. Pick the process documentation that an auditor, a buyer, or a successor would read first. Trying to fix every document at once is a project that never ends.
The Question Worth Asking Before Someone Else Asks It
Operations, quality, or compliance may hold the pen, but leadership answers for whether the work runs the way the documents say. Choose three of your most important documented processes. If a capable stranger followed each one exactly tomorrow, how much of what happened would match what you expect? Could another qualified person do the work without the informal knowledge of whoever does it now?
If you can't answer with confidence, that is your answer.
The goal is not perfect documentation. No practical process anticipates every exception. The goal is reasonable confidence that you understand how your essential work gets done and that changes reach the page. Someone will compare the page to the practice eventually. The only choice you have is whether it's you, and when.
When the Page and the Practice Part Ways
Gaps between how a company says it works and how it actually works tend to surface at inflection points such as growth, a transaction, a leadership change, or mounting performance pressure. Aspirations Consulting Group works with executives and other leaders in those moments, helping them see where documented processes, decisions, and daily practice have separated and what to do about it across the intersecting challenges that follow.
Keep Your Own Thinking Current Between Reviews
For practical insight on leadership, operations, and strategy every weekday, request a complimentary subscription to ACG Strategic Insights™, read by more than 10 million current and aspiring executives worldwide.
Thanks for reading!
~ Jerry Justice
Living to Serve, Serving to Lead™
Comments