Put the question of recorded operational decisions to a support or managed-services team and the reply tends to be one of these: “we have tickets” or “we all know why we did it”. Both answers contain a half-truth, and both explain why the same debates keep resurfacing every quarter. A decision journal, a short dated record of a choice, its alternatives, its reasoning and an expected result, is one of the cheapest tools available to operations teams, yet it is wrapped in misconceptions. Let us test six of the most common myths against what we know about how teams actually work.
Myth 1: “Our ticket system already records our decisions”
Tickets record events: who raised what, what was done and when it was closed. They are excellent for traceability of individual incidents. They are poor at capturing why a standing policy exists, such as why a particular alert routes to one rota rather than another, or why a patch window falls on a given night. Reasoning is scattered across comment threads, if it exists at all. Worse, tickets are organised by time, while decisions are needed by topic. A journal is organised the other way round, so that anyone asking “why is the escalation path like this?” can find a single page.
Verdict: tickets and journals complement each other. Link the journal entry from the relevant ticket and the gap closes.
Myth 2: “Everyone on the team remembers why”
This is true for a short time and false thereafter. Teams in support and managed services experience turnover, rotas change and incidents consume attention. The person who argued for a decision is often not the one who inherits its consequences. Hindsight bias is well documented, and it means recall of past reasoning is reshaped by known outcomes: if a decision turned out well, people remember it as having been obviously right; if badly, as having been plainly flawed. Neither is a faithful account of what was known at the time.
Verdict: collective memory degrades quickly and in a biased direction. Dated writing resists both problems.
Myth 3: “Writing it down slows us down”
Yes, there is a cost, though a modest one. A useful entry takes three to five minutes if you restrict yourself to five fields: situation, options, choice with reasoning, expected result, review date. The slowdown myth tends to assume a long design document. Compare the cost with the time lost in a recurring argument that nobody can settle, or an engineer spending an afternoon working out whether a configuration can be safely changed. Over a year, a handful of avoided archaeology sessions repays the investment many times over. Documents on core principles for effective software delivery repeatedly stress shared understanding as a driver of flow; a journal is one practical way to create it.
Verdict: the cost is minutes per decision, offset by savings that are hard to see because the avoided waste never appears on a report.
Myth 4: “A journal is only for big architecture choices”
Large choices deserve entries, certainly. But many of the most valuable entries in operations concern small standing rules: how long to wait before escalating, what threshold triggers a page, which customers get a callback. These rules look trivial yet shape daily work, and they are often the ones that nobody can explain a year later. The shift from manual operations towards code-driven practice, described in the article about why ops engineers should go from codified infrastructure to site reliability work, makes the point indirectly: once rules are expressed as code, each one embeds a decision whose reasoning should travel with it.
Verdict: use a size filter based on reversibility and dependence, not on how grand a choice feels.
Myth 5: “If a decision turns out badly, the journal will be used against us”
This fear is common and understandable, particularly in teams working under service commitments. The reality, when entries are written well, is the opposite. A good entry shows what was known at the time, the options considered and the reasoning followed. If the outcome was poor despite sound reasoning, the record protects the people involved. If the reasoning was weak, it directs improvement towards process rather than blame. Teams that adopt a blameless approach to reviews usually find the journal fits naturally, because both depend on separating decision quality from outcome.
Verdict: journals reduce blame by replacing guesses about intent with a dated record of it. Culture still matters: leaders must treat entries as learning material.
Myth 6: “A journal is just paperwork that nobody reads”
It will be, unless it is wired into existing routines. Three small connections make a difference. Reference the entry in the pull request or change request that implements the decision. Add a short review slot to a meeting that already happens. And let new team members start by reading the five most recent entries as part of onboarding. Entries that are read get corrected, extended and trusted. Entries that sit unread become clutter, and the team is right to ignore them.
Verdict: readership is a design problem, not a law of nature. Put the journal where people already look.
A note from beyond the operations room
It helps to remember that ordinary digital services lean on recorded choices as well. MotoShareGo, described as a motorbike and scooter marketplace for renting and sharing with sign-in handled by a central account service, is a case in point: any such platform exists because somebody decided how accounts, listings and users should fit together, and keeping that reasoning visible is just good housekeeping.
Limits of the argument
Promising guaranteed results would be dishonest, since nothing can prove it for your particular team. A narrower claim holds up better. Recorded reasoning is available later, memory-based reasoning is not. Recorded predictions allow calibration, and calibration is a learnable skill. Recorded options reduce repeat debates. Whether those benefits outweigh the effort for you is an empirical question, and the journal itself can help answer it: write an entry predicting how many repeat debates you will have next quarter, and review it.
One scenario to test the idea
Imagine a team that changes its out-of-hours escalation rule so that a second responder is paged after a fixed wait. The entry records the reason, perhaps a run of slow acknowledgements, plus the expected effect on response times and the extra load on the rota. At review the team finds acknowledgements did improve, but the second responder was woken more often than predicted. The entry is not a verdict on the rule. It is a precise description of the trade-off, which lets the team tune the wait rather than argue about whether to keep the rule at all.
A minimal starter for a support team
- Create a shared folder named “decisions” beside your runbooks.
- Agree on the five fields and keep entries under a screen.
- Choose a recurring monthly slot for review.
- Write entries only for hard-to-reverse choices or standing rules.
- Append outcomes under the original; never rewrite.
With those five steps, most of the myths above lose their grip. Start with the next standing rule you change, and see what the page tells you a quarter from now.