Article
How to Write a Decision Log (+ Free Decision Log Template)
Get a free decision log template with a practical example. Learn which fields to track, how ADRs and one-way doors shape it, and how to keep it current.
Teams make hundreds of decisions each quarter. Often, they disappear into Slack threads and meeting notes. Six months later, when someone asks why things were done in a particular way, nobody remembers. A decision log fixes this problem by giving every important decision a short, consistent, and searchable record.
At NuHQ, we’ve logged our decisions from the start, and they live in our knowledge base, where anyone can search them. That way, everyone can see the decisions that affect their work and the reasoning behind them.
In this guide, we’ll walk through a full decision log template, a practical example, and an explanation of each field.
What is a decision log?
A decision log is a dated, running record of a team’s significant decisions, with the context, rationale, and status of each. Sometimes, it’s called a decision register or decision record.
The idea has deep roots in software. In 2011, Michael Nygard proposed architecture decision records (ADRs): short documents with a title, context, decision, status, and consequences. Today, the ADR community site collects templates and tools for the practice. A decision log uses the same discipline for product, design, go-to-market, and operations decisions, not just architecture.
Why should a team keep a decision log?
A decision log preserves the rationale behind a team or company’s decisions, speeds up onboarding, and prevents the same debates from restarting. It turns single decisions into shared and lasting knowledge.
The practical benefits of decision logs include:
- Fewer repeated debates. When a question resurfaces, you share the entry instead of reopening the debate.
- Accelerated onboarding. New hires can read 20 entries and understand why the product looks the way it does and the context behind specific decisions.
- More effective decisions. Writing down options and evidence forces clear thinking and decisiveness.
- Accountability without blame. A record of what was known at the time keeps the decision in context.
A decision log is also one of the foundations of a single source of truth for your company.
Decision log template (with a practical example)
To create a decision log, copy the fields below into a spreadsheet, wiki, or doc. Each field should be a column, and each decision should take up a row. Required fields are noted.
Here’s a hypothetical example: a mid-market software product team is deciding how to respond to a recurring customer request to share reports.
| Field | What to write | Example |
|---|---|---|
| ID (required) | A permanent reference, like DEC-### | DEC-042 |
| Date (required) | When the decision was made (YYYY-MM-DD) | 2026-11-12 |
| Decision (required) | A short noun phrase | Read-only share links for reports, instead of a CSV export |
| Status (required) | Proposed, Accepted, Rejected, Deprecated, or Superseded | Accepted |
| Context (required) | The problem and the forces at play | Customers keep asking for an “export,” but interviews show the real need is sharing results with finance and executives who don’t have a seat on our platform |
| Options considered | 2–4 options, including “do nothing” | A) CSV export; B) read-only share links; C) scheduled PDF emails; D) do nothing |
| Decision and rationale (required) | What you chose and why, in active voice | We’ll build B. It shares information with key individuals without a seat and keeps data current. The CSV export option is deferred, not rejected |
| Evidence | Customer quotes, data, and research, with links | 38 accounts (about $520K in ARR) tagged “Can’t share results with non-users”; 6 of 8 interviewees named finance as the top recipient |
| Reversibility | One-way door or two-way door | Two-way door (CSV export can still be added later) |
| Decider (required) | One named person | Head of Product |
| Consulted | Names or teams whose input was sought | Engineering lead, Security team, two account executives, Support lead |
| Consequences | Expected results, trade-offs, and risks | 3-4 week build; needs link expiry and access controls |
| Review date | When to check whether the decision still holds | 2027-02-15 |
| Supersedes / superseded by | Links to related decisions | Supersedes DEC-031 (“Build CSV export in Q1”) |
| Links | Tickets, specs, ADRs, and feedback themes | Epic RPT-210; interview synthesis; feedback theme |
What does each field in the decision log mean?
Each field in the decision log answers one question someone will likely ask later:
- ID. A unique, permanent reference, so that tickets and docs can link to it.
- Date. When the decision was made, not when it was written down.
- Decision. A short noun phrase, as Nygard recommends for ADR titles. Make it easy to scan.
- Status. The outcome of the idea or decision. Keep reversed decisions and mark them superseded, rather than delete them. That way, you’ll maintain a full history of decision making.
- Context. The problem and the causes, whether they’re customer, technical, commercial, or political.
- Options considered. Two to four real alternatives, while always including “do nothing.” This shows the decision was a real choice versus not acting.
- Decision and rationale. What you chose and why, using an active voice: “We will…”
- Evidence. The customer feedback, data, or research behind the decision. Our guide to customer evidence explains what matters.
- Reversibility. Whether this is a one-way or two-way door (explained below).
- Decider. One named person accountable for the call.
- Consulted. The people whose input helped inform this decision.
- Consequences. Expected results, trade-offs, and risks – both positive and negative.
- Review date. When to check whether the decision still holds.
- Supersedes / superseded by. Links that make the history traceable.
- Links. Tickets, epics, specs, ADRs, and feedback themes.
How do one-way and two-way doors change what you log?
One-way and two-way doors describe how reversible a decision is. As a result, it should set how much effort each log entry gets. For example, irreversible decisions get full entries, and reversible ones get shorter ones.
Jeff Bezos described this distinction in an Amazon shareholder letter, which was summarized by Farnam Street. Type 1 decisions are one-way doors that “must be made methodically, carefully, slowly.” Type 2 decisions are two-way doors that “can and should be made quickly by high judgment individuals or small groups.” His warning was that as organizations grow, they tend to apply heavy Type 1 process to Type 2 decisions, which slows them down.
- One-way doors include a pricing model change, a data platform choice, or retiring a product line: fill in every field, consult widely, and set a review date.
- Two-way doors include a default setting, onboarding copy, or a small feature scope: fill in the required fields only within about five minutes.
Reversibility sets the minimum effort required, not necessarily the maximum. A two-way door that touches many teams might still need a full entry.
How do you decide who decides?
Assign one accountable decider per entry and record the specific people who were consulted and informed separately. Frameworks such as DACI make this explicit.
The Atlassian DACI framework splits roles into Driver, Approver, Contributors, and Informed. You can map them onto the log: the Approver is the Decider, Contributors are Consulted, and the Informed list is who gets notified when the entry is accepted and the decision is made.
How do you write a strong decision log entry?
It helps to update your decision log when the information is fresh. So, write the entry when the decision is made, keep it short, and link it from the work it affects. A strong entry should take you between 10-15 minutes.
Here’s how to approach it:
- Draft the context first. If you can’t describe the problem in three sentences, you’re not ready to decide on it.
- List real options. Include “do nothing” and at least one option that you rejected.
- Attach evidence. Link to the feedback themes or data, rather than paraphrasing them. Our guide on how to organize customer feedback shows how to make that evidence easy to link to your decision log entry.
- Name the decider and status. Mark the entry Proposed until the Approver accepts it.
- Link it everywhere. Add the decision ID to the epic, the spec, and any related ADR.
- Set a review date for one-way doors, and add it to a shared calendar.
What are common decision log mistakes?
The most common mistakes include logging too much, logging too late, and letting decision statuses go stale. Each one of these issues makes the log less trusted. When that happens, the log stops getting used.
More context on each of these common mistakes:
- Logging everything. Not every decision or choice needs an entry. Log decisions that affect customers, other teams, money, or architecture.
- Writing it weeks later. It’s easy to forget why decisions were made and who was involved. So, log your decisions within 24 hours.
- Deleting reversed decisions. Keep the full history and mark them superseded. Then link to the replacement.
- No evidence. An entry based on vibes won’t help anyone in future months or years.
- Scattered locations. Centralize your decisions in one log. One main log is better than maintaining five partial logs.
Where does NuHQ fit for decision logs?
NuHQ is an AI-native company HQ that connects knowledge, customer evidence, product decisions, planning, and delivery. In NuHQ Products, decisions link to the customer evidence behind them and the delivery work that follows.
That means the “Evidence” and “Links” fields in this template become live connections rather than pasted URLs. Also, because NuHQ includes NuHQ Knowledge, any team members searching the company’s knowledge base can find decisions, not just the product team. NuHQ’s AI summarizes research with citations back to the source.
See decisions and evidence in one place
Take the NuHQ Products tour to see how decisions, customer evidence, and delivery stay connected, or join the waitlist for the private beta.
Key takeaways
- A decision log is a dated record of significant decisions, including the context, options, rationale, evidence, decider, and status.
- Start from Michael Nygard’s ADR fields, then add evidence, reversibility, and a review date for product decisions.
- Match your team’s effort to reversibility: write full entries for one-way doors, and short ones for two-way doors.
- Never delete a reversed decision. Instead, mark it superseded and link to the replacement.
- Keep one main log, and link decision IDs from the work they help shape.
Put this into practice with NuHQ Products
Product work linked to the decisions and knowledge behind it. The beta is opening to teams in waves.
Frequently asked questions
What is a decision log?
A decision log is an ongoing, dated record of the important decisions a team has made. Each entry captures what was decided, the context, the options considered, the rationale, the evidence, who decided, and its current status. A decision log lets anyone, including people who join the company later on, understand key decisions without relying on someone else’s memory or meeting notes.
What should a decision log template include?
A useful decision log template includes: An IDDateDecision titleStatusContextOptions consideredThe decision and rationaleSupporting evidenceReversibility (one-way or two-way door)Decider and people consultedConsequencesA review dateLinks to any decision it supersedes Try to keep the required fields to about seven, so that entries take 10 minutes, not an hour.
What is the difference between a decision log and an ADR?
An architecture decision record (ADR) is a short document for one significant technical decision, a format that was popularized by Michael Nygard in 2011. A decision log is the index of many decisions across product, design, engineering, and operations teams. Many of these teams keep ADRs as the detailed entries and a decision log as the summary view that links to them. The log is the shareable, digestible artifact for the rest of your organization.
Who should own the decision log?
Each decision is owned by the person accountable for it, which is often the Approver in the DACI framework. The log itself needs one curator, usually a product operations lead, chief of staff, or engineering manager, who confirms that entries are complete, statuses are current, and superseded decisions point to their replacements. Shared ownership with no curator usually means that the log goes stale.
Where should I keep a decision log?
Keep your decision log where the work happens and where people already search: a shared spreadsheet, a wiki page, a folder of ADRs in your code repository, or a connected workspace like NuHQ that links decisions to customer evidence and work. What matters most is that you use one main location, a consistent format, and links from roadmap items and tickets back to the decisions behind them.



