Guides3 min read
The pre-launch checklist we run before every go-live
Launch day problems are almost always boring ones: a missing redirect, an unmonitored job, a backup nobody tested. This is the list that catches them.
ALBy the Adviqlabs team
Launches rarely fail because of the hard parts. They fail because of small, boring things nobody checked: a DNS record, an email that lands in spam, an environment variable that exists in staging but not in production.
This is the checklist we work through before any product goes live. Copy it and adapt it to yours.
Infrastructure
- Production environment is separate from staging, with its own database and secrets
- Environment variables are validated at startup
- Domain, DNS and SSL certificates are configured, and auto-renewal is on
-
wwwand non-wwwredirect to one canonical domain - Database has automated backups, and a restore has been tested
- Scheduled jobs and queues are running in production, not just staging
Monitoring
- Error tracking is connected and receiving events from production
- Uptime checks cover the homepage, login and a key API endpoint
- Alerts go to a channel someone actually watches
- Logs are searchable and retained for a sensible period
Security
- Every server endpoint and action checks authentication and authorisation
- Rate limiting on login, sign-up, password reset and any expensive endpoint
- Secrets are out of the repository and rotated if they were ever shared
- Dependencies are up to date, with no known critical vulnerabilities
- Admin access follows least privilege, with two-factor authentication enabled
- Security headers are set (HTTPS only, sensible content security policy)
- SPF, DKIM and DMARC records are set for the sending domain
- Transactional emails (verification, password reset, receipts) are tested end to end
- Emails render properly on mobile and in dark mode
Payments
- Live keys are in production and test keys are not
- Webhooks are registered for production and verified by signature
- Successful payment, failed payment, refund and cancellation flows all tested
- Invoices and receipts include the correct business and tax details
Product and UX
- Core loop works end to end on a clean account
- Empty states, loading states and error states are designed, not default
- Tested on real phones, not just a resized desktop browser
- Forms have validation and clear error messages
- Accessibility basics: keyboard navigation, focus states, alt text, colour contrast
SEO and sharing
- Page titles and descriptions are unique and meaningful
- Open Graph images display correctly when links are shared
-
sitemap.xmlandrobots.txtare live, and staging is blocked from indexing - Analytics are installed with consent handled correctly
Legal
- Privacy policy and terms are published and linked in the footer
- Cookie consent is in place if you use non-essential cookies
- Data processing matches what your privacy policy says
People and process
- A written runbook: how to deploy, roll back and restore from backup
- Clear owner for launch day and the first week of support
- A way for users to report problems, and someone assigned to read it
- Launch during working hours, never on a Friday evening
The first week after launch
Launching is the start of learning. In week one, we watch error rates, sign-up funnels and support messages daily, and ship small fixes quickly. Most of the improvements that matter come from what real users do in those first days.
Getting ready to launch? We can run this checklist with you, or take care of launch and support end to end. Get in touch.