When It Breaks

Status page templates and on-call guides are built for the engineering response. They stop where your job starts.

When something breaks, support and CS are the ones answering the customers. You need to spot the pattern in the queue before anyone declares an incident, reply to fifty people without guessing at the cause, call your biggest accounts before they hear it from their own staff, answer the credit request fairly, write the letter afterwards, and handle the customer executive who emails your CEO.

When It Breaks covers that side, in 24 plays. Each one gives you:

• what tends to go wrong in that moment, and why
• the move, step by step, and who does it
• the words to say on the call and in writing
• if-then decision rules
• the mistakes that make things worse
• an agent prompt that gathers facts from your tickets and incident channel and drafts the message while you handle the people
• a check step for the agent's draft, so a guessed cause or fix time never goes out

Say what you know. Give a time for the next update and keep it. Never promise what you can't keep. And where you need a number, like how often to update or how much goodwill to propose, you get a labelled starting default with the reasoning shown.

What's inside

• 24 plays in 5 chapters: the first hour, keeping people informed while it's broken, after it's fixed, executive escalations, and learning from it
• A which-play map and an outage timeline from support's side: who does what, and what goes out when
• One made-up incident followed from first ticket to review, plus adjustments for messy incidents: partial or regional, slow not down, supplier-caused, possible security or data incidents, and out of hours
• Decision tools: starting defaults for update cadence and for credits and goodwill, two worked credit calculations with made-up contract terms, and four quick decisions (raise it or wait, call or email, credit or goodwill, close or hold)
• Four judgement calls worked through, common advice set against what to do instead, a vague update beside the fixed one, and bad agent drafts with each error marked
• A made-up exec escalation from first email to close, and two made-up incidents that went badly
• A reopen and complaint signals worksheet for every incident
• 12 templates: a seven-part macro set built around what the customer is doing, a big-account call script and follow-up email, an internal incident FAQ, a support-to-incident note, an after-incident letter and a filled example, credit request replies (due, not due, goodwill), exec complaint first reply, plan and close, a support seat card, a queue math sheet, a status page and macro sync check, and a post-incident support checklist
• Three support seats that make incidents calmer: writer, queue keeper and account caller
• Policy check boxes wherever cause statements, data, credits or contracts come up, sending you to your own legal, finance and security owners
• A 47-page PDF, plus a free sample with 3 complete plays (The holding reply, Big accounts hear it from you, Credit and SLA requests)

Who it's for

Support leads, CSMs and account owners at software companies who face customers during outages, major bugs and executive escalations.

Who it's not for

Engineering incident commanders looking for technical response process, or anyone looking for legal, security or financial advice.

The fine print

Written by Draft First, a customer success and support lead. The companies, people and incidents in the examples are made up. Where the book gives a number, it's a labelled starting default with its reasoning, not a published target. Not legal, security or financial advice.

Questions: hello@draftfirstcs.com

All sales are final.

Get When It Breaks

Questions before you buy? Email hello@draftfirstcs.com.