Website Maintenance
The unglamorous work that keeps a site from quietly degrading: patches applied, backups tested, uptime watched, and someone who answers when something breaks.
The unglamorous work that keeps a site from quietly degrading: patches applied, backups tested, uptime watched, and someone who answers when something breaks.
Sites rarely fail dramatically. They drift — a plugin update breaks a form and nobody notices for a month, an SSL certificate lapses, images get uploaded at full resolution until the homepage weighs nine megabytes, a redirect gets lost in a migration and a page that used to rank stops existing.
Maintenance is largely about catching those early. We monitor uptime, Core Web Vitals, broken links, form submissions and search visibility, and we tell you what changed rather than only reacting when you notice.
Everything goes to staging first, gets checked against the paths that matter — forms, checkout, login — and only then goes live, with a rollback ready. Updating straight to production is how a Tuesday afternoon becomes an incident.
On staging first, with a rollback path.
Patching, malware scanning, hardening and access review.
Off-site, automated, and periodically test-restored.
Alerts to us, not just to your inbox.
Core Web Vitals tracked over time, not once.
A monthly allowance of small changes.
Updates on staging, security patching, off-site backups with tested restores, uptime and error monitoring, a monthly performance and search-visibility check, and an allowance of small content or fix requests. Larger work is quoted separately so the plan stays honest.
Yes, after an audit. We need to know what we are taking responsibility for — hosting, code quality, dependency state, backup status — and occasionally the audit concludes that some remedial work has to happen before a plan makes sense.
Response times are set in the agreement and depend on severity. A site down or a checkout failing is treated differently from a typo. We would rather commit to something we can hold to than advertise a number we cannot.
Take it offline or into maintenance mode, restore from a known-good backup, find and close the entry point, rotate credentials, then request a review from Google if it was flagged. Having tested backups is what makes this hours instead of days.