Turn a messy decision into a record worth keeping

# Example decision register — Ops decisions

*Fictional company: a 40-person distributor. Three entries, newest first. Together they show: an everyday four-line entry, a deep entry, and what supersession looks like — the point being that the trigger worked, not that anyone was wrong.*

---

#### 2026-09-04 — Helpdesk tool approved; supersedes 2026-03-12 inbox decision

- **Decision:** Buy the helpdesk tool (basic tier, 3 seats), migrate the shared inbox by end of October.
- **Owner:** Dana (ops manager)
- **Why it made sense:** The 2026-03-12 "Reopen if" fired — weekly requests have exceeded 40 for five consecutive weeks (reseller program traffic, visible in the monthly counts since July). The original decision's reasoning was volume-based; the volume changed. This is the trigger doing its job, not a reversal of judgment.
- **Reopen if:** Usage at renewal shows we're paying for seats we don't use — reopen the tool choice, not the need.
- **What changed since the original decision:** The reseller program added ~25 requests/week — the June unknown, now resolved.
- **Actual approver:** Priya (COO)
- **Supersedes:** 2026-03-12 — "Support requests stay in the shared inbox"
- **Source:** Ops/finance monthly review, 2026-09-04

---

#### 2026-05-02 — New-customer welcome calls stay with account managers, not support

- **Decision:** Welcome calls for new accounts stay with the assigned account manager. Support doesn't take them over.
- **Owner:** Dana (ops manager)
- **Why it made sense:** The AM already knows the account from the sales handoff; a second introducer adds a handoff without adding value. The support team's pitch to take them was about their capacity being visible, not about the customer.
- **Reopen if:** AM headcount drops below four, or onboarding satisfaction dips two quarters running.

*(An everyday entry. Four lines, two minutes, done. Most of your register should look like this.)*

---

#### 2026-03-12 — Support requests stay in the shared inbox; no helpdesk tool this year

- **Decision:** Customer support requests continue running through the shared `support@` inbox with the tagging convention. We are not buying a helpdesk tool this year.
- **Owner:** Dana (ops manager)
- **Why it made sense:** Volume was ~22 requests/week, handled by two people who share context daily. Every helpdesk tool quoted came with per-seat pricing that costs more than the problem: the pain was occasional dropped follow-ups, not volume.
- **Reopen if:** Weekly requests exceed 40 for four consecutive weeks, or a third support hire is approved — whichever comes first.
- **Actual approver:** Priya (COO) — the budget call was hers
- **Facts known:** 22 requests/week average over the last 8 weeks (inbox count). Two handlers. Three dropped follow-ups in Q1 — all recovered, one customer annoyed.
- **Assumptions:** Volume stays flat through the fall — *confirmed or denied by the monthly inbox count.* The tagging convention is actually being used — *spot-check in April.*
- **Unknowns:** Whether the new reseller program (launching June) changes request volume. Accepted: we'll see it in the monthly count long before it breaks anything.
- **Alternatives considered:** Helpdesk tool — rejected on cost at current volume. A rota rule for follow-ups inside the inbox — adopted as part of this decision, not instead of it.
- **Risks accepted:** If volume spikes fast, the inbox breaks before a tool can be bought and set up.
- **Source:** Ops/finance monthly review, 2026-03-12 (meeting notes in drive)
- **Superseded →** 2026-09-04 entry *(appended 2026-09-04; original text unchanged)*

---

## What to notice

- **The March decision was right, and it still got superseded.** Good decisions expire when their conditions change — "Reopen if" is how you notice. The June unknown (reseller program) was accepted with a stated reason, and it's the exact thing that later fired the trigger. Nobody had to reconstruct that from memory.
- **The May entry is the everyday shape.** Four lines. No status, no source, no depth — none of it was needed.
- **Depth showed up only where stakes did.** The March entry carried facts, assumptions, and unknowns because a budget call rode on them.
- **The September entry names what changed.** No delta, no reopening.
- **The March entry's body was never edited.** One "Superseded →" line was appended. New entries don't carry a status at all — a decision is active until a line like that appears.

---

## From the maker's own register

*A real entry from building this kit — included here so you can see the format holding a real decision (and so the blank template stays blank):*

#### 2026-08-18 — This kit ships without a spreadsheet version

- **Decision:** The decision log ships in Markdown, Google Doc, and Notion formats only. No Google Sheet or Excel version.
- **Owner:** Chelsie Hodgkiss
- **Why it made sense:** "Why it made sense" is the whole point of this artifact, and it dies in a spreadsheet cell. A one-line grid row is exactly the decision log this one exists to replace.
- **Reopen if:** Written requests for a spreadsheet version reach five.