Blog / Case studies /
Nine months of client websites: what I’ve learned
Nothing on a client website stays in its own box. A content change touches a template. A plugin update breaks a form. A DNS slip makes a perfectly healthy site look dead.
That’s the biggest thing nine months of this work has taught me. Good website management is less about heroic fixes and more about knowing what depends on what, telling people clearly, and leaving notes the next person can use.
I joined Nivonix in October 2023 and now work across WordPress, Shopify and other CMS and e-commerce sites. I can’t share client names or private outcomes, but the patterns repeat. Every site has its own quirks, and the safest process starts with context.
Learn the site before changing it
A request that sounds simple can involve a theme, a plugin, a hosting setting and an outside service. So before I change anything, I map the pieces: domain, DNS, hosting, CMS, theme, key plugins and integrations. Who owns each account? Where does the backup live?
Then I ask what the change should achieve and who’ll use the result. “Make the page faster” needs a page, a user problem and a way to measure. “Fix the email” needs me to separate form submission from email delivery. Clear questions stop you fixing the wrong layer.
Keep a change record
Record the request, what was changed, where it was changed and how it was tested. A short note can save time when another issue appears later. It also makes handovers less dependent on one person remembering an admin screen from three months ago.
You don’t need an elaborate ticketing system. A shared log with date, site, change, test and follow-up is enough if you keep it current. Keep credentials in an approved password manager, not in the same record.
Test the customer journey
A site can load and still fail at the one thing that matters. After a change, test forms, booking links, checkout, email notifications and account flows. If there’s a staging copy, use it for anything that could touch templates or payment paths. For low-risk changes on a small site, a backup and a planned window can be more practical, but know the recovery route first.
I always test from the public site, the way a visitor would. A successful save in the admin doesn’t prove a customer sees the new page or gets the notification. Keep evidence of the result, especially for recurring workflows.
Treat updates as managed work
Updates keep software maintained, but they can break integrations and layouts. Read the release notes, apply updates in a controlled way and check the important functions afterwards. If something breaks, stop. Don’t change three more things; isolate the cause and use the rollback.
The right cadence depends on risk. Security fixes need quick attention, and the rest can wait for a scheduled window. Leaving updates forever isn’t a plan. Neither is applying them all untested.
Explain the trade-offs
Website work often involves constraints. A theme may be easy to edit but load extra assets. A custom feature may be flexible but need more maintenance. A migration may improve control but require careful URL mapping. Present the options, risks and required approvals in plain language so the business can make the decision.
Clients don’t need every implementation detail. They do need to know what’s changing, what could be affected and what they have to do. Communication is part of delivery, not a courtesy added afterwards.
Build knowledge that survives handover
Document where important settings live, how routine checks happen and who to contact when a service fails. Train the person who will own the site after a project closes. A handover that only transfers passwords leaves the next person guessing.
So I try to think about the next person while I’m doing the current job. Will they understand this custom setting? Is the change reversible? Does the client know how to update content without breaking the layout?
FAQ
What is the hardest part of managing several websites?
Keeping context and ownership clear across different platforms and integrations. A shared process helps, but each site’s critical functions still need individual attention.
Should every change be tested on staging?
Not every small edit needs a full staging workflow, but changes with technical or commercial risk should be tested in a safe environment where possible.
What should a website handover include?
Account ownership, access method, hosting and domain details, backups, important integrations, maintenance routines and support contacts.
Leave a record for the next person
The habit that saves the most confusion is recording not only the change, but the reason for it. If a plugin is disabled because it conflicted with another feature, note the pages tested and the behaviour that should remain. If a DNS record is changed, include the system where it was changed and who owns the account, but keep credentials in the approved password manager. A handover should be useful without exposing secrets.
This matters most when several people are involved. The owner knows why a page exists, a designer understands the theme, a developer knows the deployment. A short note lets those pieces meet. It also helps a future maintainer tell an intentional setting from an accident. Good maintenance isn’t about being the only person who knows the answer. It’s about making the answer findable.
Related reading
- Training clients to manage their own websites
- Client website handover documents: what I include
- Nine months of client websites: what I’ve learned
Pick one site you manage and write down its critical dependencies and recovery route. Could someone else follow your notes? If not, fix that before the next urgent change.