Blog / Web management /
A clear website escalation guide for non-technical teams
A checkout button stops responding. The person who spots it sends one message: “The website is broken.” The technical team has to ask which page, which device, what happened and when. Meanwhile, customers may still be trying to buy.
An escalation guide helps people report a problem so the right person responds. It shouldn’t expect non-technical colleagues to diagnose anything. It should help them recognise urgency, gather safe evidence, contact the right owner and avoid making things worse.
Define what counts as urgent
Start with impact, not technical vocabulary. Can customers buy? Is the site unavailable? Are forms failing? Is published information materially wrong? Has the team noticed suspicious activity or an unexpected change? These are easier to recognise than an error code, and they help route the problem.
Set clear severity levels with examples relevant to the organisation. A critical issue may involve a complete outage or a checkout failure affecting customers. A high-priority issue may affect an important workflow or a large section of the site. A lower-priority issue might be a visual defect with a workaround. The definitions should reflect real business impact and the actual support agreement. Never invent response-time promises.
Explain what to do if the impact is unclear. A report can say, “I cannot confirm whether this affects everyone.” Give them a safe contact route for uncertainty. Don’t force anyone to pick a severity they don’t understand.
Give people a short evidence checklist
Ask the reporter to capture details that help reproduce the issue: exact page URL, approximate time and time zone, the action they took, what they expected, what happened instead, device and browser if known, and whether the issue repeats. A screenshot or short screen recording may help, as long as it does not expose customer information or credentials.
Ask for a plain description, not a diagnosis. “The save button spins and the page doesn’t confirm the change” beats “the database is down” when nobody has seen the database. When I pick up a site, I’d much rather get that first line than a guess. A screenshot of an error helps, but strip out tokens and personal data first.
For a customer report, capture only the information needed to investigate and follow privacy rules. Don’t ask staff to forward whole customer accounts or payment details in a general chat. Use approved support systems with proper access controls.
Route the issue to a named owner
List the contact route for each type of problem: internal web owner, hosting provider, commerce platform, payment provider, security contact or content approver. Provide the method that is actually monitored, such as a support portal or an on-call number where one exists. A list of names with no sign of who’s available won’t help in an incident.
Clarify when to contact an external provider directly and when an internal owner must coordinate. The company may have an agreement that requires a ticket from a designated account holder. A colleague who opens a second ticket with no context can split the investigation in two. Say who owns the relationship and what to include.
Include a backup route if the primary contact is unavailable. Avoid promising a response time unless it is part of a verified support arrangement. If no after-hours support exists, state the real process and define what staff should do to preserve customer safety and evidence in the meantime.
Tell staff what not to change
People often try to help by editing the page, clearing settings or deactivating a plugin. During a live problem, those actions can wipe out evidence or make things bigger. I’d list the changes that need an authorised owner: DNS updates, payment settings, plugin changes, theme deployments and security controls.
Never ask staff to ignore a security warning, share a password in chat, install remote-access software or paste secret keys into a ticket. If you suspect a compromise, contact the security or technical owner through the approved route and keep the details. The guide should fit the organisation’s incident process, not compete with it.
For low-risk content corrections, the process may allow trained editors to act. Make the boundary clear: which pages can they update, who approves claims and how should a change be logged? A sensible permission model gives colleagues room to work while protecting critical settings.
Keep communication calm and useful
When an issue is active, nominate someone to send updates. They can tell affected teams what is known, what is being checked and when another update is expected, if that timing is confirmed. Avoid speculative explanations. “We are checking the payment flow” is more honest than blaming a provider before the cause is known.
Use one record for key decisions and timestamps. This reduces repeated questions and helps the next person understand what has already been tried. After resolution, record the cause if known, the action taken, any remaining risk and whether a follow-up is needed. If the cause is unknown, say so.
A post-incident review should focus on improving the system, not finding someone to blame. Did the guide route the problem correctly? Was evidence easy to gather? Did staff know which actions were safe? Update the process where it failed, and tell the affected teams what changed.
Make the guide easy to find and use
Keep the first page short. Put urgent contact details and first actions at the top. Add deeper instructions for common issue types below, with links to the relevant runbook. Use plain words and make the guide accessible to someone who is worried and in a hurry.
Test the document with a colleague who did not help write it. Give them a scenario, such as a product page that shows the wrong availability or a form that fails. Can they determine who to contact, what information to include and what not to change? Revise anything they have to guess.
Review the guide after staffing changes, platform migrations, support-contract changes or a major incident. A stale contact list is worse than no process, because it suggests help is available when it isn’t. Assign an owner and a review date.
Make the first report better
You’re not trying to turn every colleague into a web technician. You’re making the first report clear enough that the right person can act, and stopping well-meaning changes from adding risk.
Related reading
- Building a website escalation guide for non-technical staff
- How to explain website downtime to non-technical stakeholders
- What to prepare before a website host changes support teams
Pick one common website problem and ask a non-technical teammate to follow your guide. Can they describe the impact, capture safe evidence and reach the right owner without guessing? If not, make that path simpler.