Contents
I do not want a decision log to make every earlier version of my thinking look wise. I want it to explain why a choice seemed reasonable, what changed and which choice is current. Without that small history, a later reader can mistake an old proposal for a live plan or assume that a changed view was simply indecision. The record should make revision easier to understand, including the times when the original account contained an error.
For this walkthrough, I will use a fictional reading group choosing between a folded handout and a single sheet. The preferences and responses are invented for the demonstration; there is no claimed test, client or measured outcome. The choice is ordinary enough to show the structure without requiring a formal project system. My proposed log can live in plain text, a shared document or a note application that preserves the fields and links the group actually needs.
Write the choice before the debate
The first line I would want is the current decision. Suppose the fictional group initially chooses a folded handout for its next discussion, subject to reviewing a sample. The log should say exactly that. Folded handout chosen without the condition would turn a provisional step into approval for production. Conversely, a page beginning with several paragraphs of debate can leave the actual choice hidden. I would put the operative sentence first, then explain the reason and the status in ordinary words.
The record also needs a scope. This decision concerns one discussion, not every future handout or the group's identity as a whole. I would name the issue being resolved: which format to prepare for review. That prevents the note becoming a general rule when someone later copies it into a planning guide. The smaller scope can be reassuring. A local choice does not have to settle the merits of every possible format in order to be worth recording.
I would then identify who made the decision through the group's actual process. A personal note can simply say my current choice. A shared record should not imply group approval if only the writer expressed a preference. In the example, the group agrees to review the folded sample, but that agreement does not assign production work to an absent volunteer. Responsibility, spending and final approval may need separate actions. A compact log can show these boundaries without becoming a transcript of the entire conversation.
Keep the reason small enough to inspect
For the imagined choice, the reason might be that the proposed discussion has two distinct parts and the group wants to see whether separate panels make the material easy to use. I would describe that as the group's design intention, not as evidence that folding will improve comprehension. A reason can be a preference, constraint or working hypothesis. Naming its type makes the choice more understandable than dressing every rationale in the language of research findings.
I would record the alternative considered. A single sheet remains an option because it may be easier to compare both parts at once. The log does not need to prove the rejected option was bad. It should explain why the group chose one route at this point. That distinction matters when a later change makes the alternative appropriate. The earlier note can then provide useful context rather than a verdict that must be defended or quietly erased to make the revised plan look consistent.
The rationale should point to relevant materials without duplicating them. If there is a sample layout, I would identify its version. If the decision relies on a source, the source record belongs close enough to find. A log containing only a link can fail when the destination is unavailable, so I would also summarize the specific basis in one or two sentences. The link provides inspection; the short explanation lets the decision remain intelligible when read outside the original working space.
State what would make me reconsider
A useful reconsideration condition is specific to the decision. For the fictional handout, it might be that reviewing the sample reveals the two parts need to be visible together. That condition differs from saying review later. It identifies the question the sample should help the group answer. I would not promise that the review will produce an objective winner. The group may still need to make a preference-based choice after seeing the alternatives, and the log can preserve that fact.
I would distinguish a trigger from a routine task. A trigger describes a circumstance in which the decision becomes relevant again. A task describes work someone has accepted now. The decision might remain valid until the text changes substantially, without requiring a volunteer to inspect it every week. Recording every possible future concern as a scheduled task can make a small project harder to run. I would retain the meaningful condition and avoid creating work merely to make the log appear thorough.
This is also where I would name the uncertainty I am accepting. Perhaps neither layout has been reviewed by the group yet. The current decision is to prepare one sample first, not to establish the final format. A clear status makes that honest. I would prefer a modest record that says what remains unknown to an elaborate confidence score with no shared definition. The log should help another person understand the risk or tradeoff, not simply display that uncertainty has been considered.
Decision D-04 — illustrative reading-group handout Scope: Format to prepare for the next discussion's review. Initial choice: Prepare a folded sample first; final production is not approved. Reason: Two discussion parts may benefit from separate panels. Alternative: Single sheet with both parts visible together. Reconsider when: Reviewing the sample shows the two parts need continuous comparison. Update 1 — constructed example, not real feedback New information: In the imagined review, the group chooses to compare the two parts side by side. Current choice: Prepare a single-sheet revision for final checking. Reason for change: The chosen discussion activity now requires simultaneous visibility. Earlier record: Retained as context; folded sample decision is superseded. Still open: Who checks the final text and who approves its use?
Add the new information without changing its type
Now imagine the group changes the activity. Instead of reading two parts in sequence, it decides to compare them side by side. That is a new design requirement created by the group, not a discovery that folding is universally ineffective. I would write the change in those terms. A common mistake is letting a local preference become a general lesson: folded handouts do not work. The log is more useful when it identifies the actual new condition that affected this particular choice.
I would add the information with its source and date where appropriate. In the constructed example, it is the group's later decision, explicitly invented here to show the process. In a real project it might come from an agreed meeting note or a documented request. I would keep a tentative suggestion separate from a confirmed requirement. If someone merely wonders whether comparison would help, the next action might be discussing the activity rather than immediately replacing the layout. The new information must earn the role the log gives it.
The current choice can then change with an explanation. Prepare a single-sheet revision is an action under the new activity, while final approval still awaits the relevant check. The former folded-sample choice becomes superseded. I would not delete it or keep presenting both as equally active. A reader needs to identify which record governs the present work. The history remains useful because it explains why an earlier sample exists and prevents another contributor from reviving it under an obsolete assumption.
Distinguish a changed decision from a corrected error
A different scenario requires different language. Suppose the original log says the group approved the folded handout, but the meeting only approved preparing a sample. The later change is a correction to the record, not a new group decision. I would say the approval wording was inaccurate and restore the actual scope. Presenting this as a change of mind would unfairly attribute a decision to the group that it never made. A small correction can carry a significant difference in responsibility.
I would identify any practical consequences. Has somebody already begun final production because of the incorrect line? Does an action list or shared calendar repeat it? The correction needs to reach those affected places. Editing the canonical note is necessary, but it may not inform a person working from an earlier attachment. I would use an appropriate short update that names the corrected scope and the action now required. The aim is coherent work, not a historical record whose accuracy nobody relying on it knows has changed.
The same distinction applies to source errors. If a log misstates a policy, I would correct the description before discussing whether the decision still makes sense. The choice may need revision, or its other reasons may still support it. Those are separate questions. Keeping error correction separate from judgment allows a project to learn without making every correction a defense of the whole decision. It also gives a future reader a clearer account of what was known, what was misunderstood and what was chosen.
Use fields to find the current state
For a small personal log I might use plain labels: decision, status, basis, alternative and next condition. A collection with several active records can benefit from consistent structured fields. Obsidian's Properties documentation describes typed fields stored at the beginning of notes. If I use a status property, I would define its meaning for this collection. Proposed, current and superseded could be local values. The application can store them; it does not decide which value is warranted by the group's agreement.
I would keep nuanced reasoning in prose. A field called reason is not useful if it compresses the whole tradeoff into an unexplained abbreviation. Nor would I use one date for creation, approval and revision without saying what it means. The date a note was typed can differ from the date the decision was made. The next reader should not need to infer that distinction from file metadata. A few clear fields paired with a short explanation can serve retrieval without forcing every record into a rigid form.
Links help connect an update to the record it supersedes. Obsidian supports internal links to notes and headings, with documented format and portability considerations. I would use a connection whose visible text explains the relationship: this replaces the folded-sample choice for the next discussion. A generic related link would leave the important meaning unstated. If the log travels outside the application, I would inspect a representative output rather than assume that application-specific navigation will remain available to every recipient.
Keep the present choice easy to find
A long history can be accurate and still difficult to use. I would place a short current-state summary above the sequence of changes. It should name the active choice, its remaining condition and the record containing its basis. The summary is an aid to navigation, not a second independent version of the decision. If it repeats details, I would update it alongside the canonical record. Otherwise the most visible sentence may remain stale while the deeper history is perfectly maintained.
For the fictional handout, the summary would say that a single-sheet revision is being prepared for final checking and no final production approval is recorded yet. That is enough for a contributor deciding which sample to edit. The older folded layout remains a historical artifact. I would store it with a clear relationship to the log rather than rename it so aggressively that its origin becomes mysterious. The record explains why it exists; the active summary explains why it should not be used as the current instruction.
I would also make disagreements legible. A group member may prefer the earlier format without disputing the current decision. The log can preserve a material objection when it affects interpretation, while still identifying the process the group used to proceed. It should not imply unanimous enthusiasm where there was only an accepted practical choice. A clear current state is compatible with an honest account of alternatives. It prevents the history becoming either a victory narrative or an indecipherable collection of equally weighted positions.
Close the record when its job ends
Once the handout is used, I would distinguish its final state from any future design question. The log can record what was approved and for which discussion. It need not remain an active decision simply because another handout may be needed someday. A future project can consult it as context and make a new choice. Reusing the record should not import its conditions automatically into a situation with different readers, content or constraints. The scope written at the beginning now helps the record end cleanly.
My closure note would mention only the information needed to interpret the outcome. If no observation about the handout's use was collected, I would not invent a successful result. Used as approved is different from proved effective. If the group later supplies a specific comment, it can be recorded with its source and limits. The decision log is not obliged to become an evaluation study. It can simply preserve the choice, its development and the point at which work under that choice was completed.
For a recurring series, I would separate a reusable convention from this local history. Perhaps the group adopts a rule that every handout gets a visible approval state. That is a new convention with its own scope, not the automatic conclusion of having maintained this record. I would write it where future work begins and link back only if the history is useful. A decision log earns its place by explaining consequential changes, then allowing the project to move on without either erasing or endlessly reopening them.
Frequently Asked Questions
Q: Should a decision log record every preference?
A: No. I would record choices whose basis or later revision matters to the work. Ordinary passing preferences may need no formal entry. The amount of history should help a future reader understand an important decision, not turn every small edit into a committee process.
Q: Is superseded the same as wrong?
A: No. A decision can have fit its original conditions and later be replaced because the conditions changed. A factual error or inaccurate record needs a correction. Use the wording that reflects what actually happened rather than labeling every change an improvement over a mistake.
Q: Can I edit the old record instead of adding an update?
A: Small clarifications may be reasonable, but preserve consequential history when others need it to interpret earlier work. Make the current choice unmistakable and explain changes that affect actions, authority or the basis of a decision. Follow any established record requirements for the organization.
Q: Do I need special decision-log software?
A: No. Plain text can hold the complete example. Structured properties and links are optional tools for retrieval when several records need maintaining. Use a format that the relevant people can read and revise through an agreed process.
Q: What if I cannot identify the original reason?
A: Say that the basis is not recorded rather than supply a plausible retrospective explanation. You can document what is known now and decide whether the current choice needs review. Honest gaps are more useful than a tidy rationale that nobody actually used.
One thing to take forward
I would keep the change in view: the earlier choice, the new condition and the decision that applies now. That small sequence lets a record explain revision without confusing it with error or indecision. The useful log does not prove that I always chose well. It gives another reader enough context to understand the choice, question its basis and recognize when a different decision has legitimately taken its place.
Sources & further reading
Something needs correcting? Write to the editors with the article link.



