How to write a one-page support playbook
A fill-in-the-blanks template for a single-page support playbook: addresses, owners, hours, triage, escalation, refunds, tone, and the links people need.

On this page(7 sections)
Most small teams don't need a support wiki. They need one page that answers the questions people ask in chat ten times a week: who handles this, how fast, and am I allowed to refund it?
A one-page playbook is short enough that people actually read it and specific enough that a new teammate can work a shift with it open in another tab. This guide gives you the template, explains what goes in each section, and ends with a filled-in example.
Why one page
Long support documentation fails in a predictable way. Nobody reads it end to end, it drifts out of date, and people go back to asking the person who has been there longest. A single page forces you to make decisions instead of describing every possible case. If a rule doesn't fit on the page, it's usually either rare enough to handle by judgment or important enough to deserve its own linked document.
The page has three jobs:
- Tell people where mail lands and who owns it.
- Set the defaults (response times, tone, what to do when stuck).
- Grant authority so people don't have to ask permission for routine decisions.
The template
Copy this into whatever your team uses for documents. Keep the headings; replace the bracketed text.
SUPPORT PLAYBOOK — [Team name] Last reviewed: [date] by [name]
1. ADDRESSES AND OWNERS
[address] → owner: [person/role] backup: [person/role]
[address] → owner: [person/role] backup: [person/role]
2. HOURS AND TARGETS
Covered hours: [days, hours, time zone]
First reply target: [e.g. same business day]
Urgent (service down, data issue): [target + how to reach on-call]
3. TRIAGE RULES
Work first: [what counts as top priority]
Then: [next category]
Can wait: [low-priority category]
Never reply to: [spam, vendor pitches, etc.]
4. ESCALATION
Bugs → [where and how]
Security or account access → [who, and what never to do]
Angry customer / legal threat → [who]
Founder/lead gets pulled in when: [criteria]
5. REFUNDS AND MONEY
You may approve without asking: [scope]
Needs approval from [person]: [scope]
Always log: [where]
6. TONE
Sign off as: [first name / team name]
We always: [2–3 habits]
We never: [2–3 habits]
7. LINKS
Docs: [url] Status page: [url] Snippets: [url]
Bug tracker: [url] Billing admin: [url]
Filling in each section
1. Addresses and owners
List every address that receives customer mail, even the ones that "barely get anything." The forgotten address is the one where a complaint sits for three weeks. Every address gets one owner and one backup. A role ("whoever is on rota") is fine as long as the rota itself is linked.
2. Hours and targets
Write targets you will actually hit on a bad week, not a good one. "Same business day" is honest for most small teams; "within one hour" usually isn't. Spell out the time zone. If you offer any after-hours coverage, say exactly what qualifies as urgent so the on-call person isn't woken for a password reset.
3. Triage rules
This is the order of work, not a full taxonomy. Three or four lines is enough. A common order is: anything blocking a paying customer, then billing, then how-to questions, then feature requests. Add a line for what never gets a reply so people stop agonizing over vendor pitches.
4. Escalation
For each type of escalation, name a destination and a format. "Tell engineering" is not a rule; "Open an issue in the Support Bugs project with repro steps and the thread link, then post the link in the support channel" is. Include at least one explicit "never" for account-access requests, such as never changing an account owner based on an email from an unverified address.
5. Refunds and money
This section removes the most friction. People hate asking permission for small decisions, and customers hate waiting while they do. Pick a threshold below which anyone can refund without approval, and say where to record it. Above the threshold, name one approver.
6. Tone
Three "always" and three "never" lines beat a page of brand-voice prose. Examples: always answer the question in the first paragraph; always give the next update time; never blame the customer; never promise roadmap dates.
7. Links
The page is a hub. Link the docs, the status page, the snippet library, the bug tracker, and whatever admin tool is used for billing. If someone has to ask "where is that?", the link belongs here.
A filled-in example
Here is the playbook for a fictional four-person SaaS company called Example Analytics. Treat the specifics as illustration, not recommendation.
| Section | Example Analytics |
|---|---|
| Addresses | support@ → Priya (backup: Sam); billing@ → Sam (backup: founder); security@ → founder only |
| Hours | Mon–Fri, 9:00–17:00 US Eastern. First reply same business day. Urgent: page on-call via the incident channel |
| Triage | 1. Paying customer blocked. 2. Billing. 3. How-to. 4. Feature requests. Vendor pitches: archive, no reply |
| Escalation | Bugs → issue in "Support Bugs" with repro steps + thread link. Account access → founder; never change owner email from an unverified request |
| Refunds | Anyone may refund up to one month of the customer's plan. Larger or repeat refunds → Sam. Log every refund in the billing notes |
| Tone | Sign with first name. Always: answer first, give next update time, thank them for detail. Never: blame, promise dates, use ALL CAPS |
| Links | Docs, status page, snippet doc, bug tracker, billing admin |
Notice what's missing: no paragraph on company values, no history of why the rules exist. If the "why" matters, link a separate note.
Keeping it alive
A playbook that was accurate in March is wrong by September. Three habits keep it honest:
- Put a "last reviewed" date and name at the top. A stale date is a visible prompt.
- Review it in your weekly or monthly support meeting. Ask: did anything happen this period that the page didn't cover?
- Change the page when you change the rule. If the refund threshold goes up in a chat thread, the page gets edited the same day.
When a new hire starts, have them read the page and then mark anything that confused them. Their questions are the best edit list you'll get.
Common mistakes
- Writing aspirations instead of rules. "We aim for delightful support" gives nobody a decision.
- No authority. If every refund needs approval, the page hasn't removed any friction.
- Too many priorities. If six things are "top priority," nothing is.
- Burying urgent paths. The on-call route should be findable in five seconds.
Bottom line
A one-page support playbook works because it is short, specific, and gives people permission to act. Start with the seven sections above, fill them with the decisions you already make informally, add a review date, and edit it whenever reality changes. If it grows past a page, move detail into linked documents and keep the page as the hub.
Start with Koltrix
Your domain, one inbox, and an API that sends.
A team inbox where AI sorts and drafts (nothing is sent without your click), plus the transactional API and SMTP relay your product sends with. 7 days free, no card.


