Skip to article content

How to prioritise website improvements: what to address first

When several website improvements are pending, establish what is going wrong, who is affected and what can wait. A worked example shows how to put them in order.

Turn the improvement list into specific decisions

One report recommends security changes, another discusses speed, and someone proposes a redesign. Everything seems important, but time and budget are limited. Separate observed problems, known consequences and unanswered questions.

Prioritising involves more than numbering a list. To put tasks in order, consider the problem each one addresses, what could happen if you leave it open and whether another task needs to happen first. Sometimes the first task is a correction; sometimes it is finding the missing information.

Gather observations in one place with their dates and sources. “The enquiry form does not deliver messages” needs different checks from “the website looks old”, even when both enter the same discussion.

What to record before deciding which improvement comes first

A decision record for each need
FieldQuestion to answer
ObservationWhat happens, and where can it be checked?
EvidenceWhich test, data or document supports what is known?
ConsequenceWhich function, information or activity is demonstrably affected?
UrgencyWhat justifies attention now: an active fault, a deadline or known exposure?
Effort and costWhat does the next step involve, and which estimates still need confirming?
UncertaintyWhat is missing before choosing without assuming a cause?
DependenciesWhich person, access or prior task is needed?
Next stepWhich action is reasonable now, and how will it be reviewed?

The matrix structures a discussion rather than producing a universal score. If impact is unknown, record that and define how to establish it. An invented risk or return figure would make the list look precise without making it useful.

Distinguish action, investigation, planning and observation

  • Act: sufficient evidence exists, consequences need attention and a change can be prepared with suitable checks.
  • Investigate: the signal matters, but cause, scope or impact needs confirmation before choosing an intervention.
  • Plan: the need is understood and requires coordination, dependencies or a broader project.
  • Observe: a sample or trend review is needed before making significant changes.

To order tasks within a group, compare the consequence of leaving them open, urgency and dependencies. Effort helps with scheduling, but an easy task does not deserve priority just because it is quick. If impact or cost is unknown, a focused check can be the next priority.

A need can move between groups. Investigation should have an owner, a question and a review point rather than become indefinite postponement. An ongoing security incident requires the responsible team’s incident procedure before ordinary improvement planning.

A worked example with four competing needs

Fictional example for practising the method. It is neither a client assessment nor measured client results.

Decisions based on the available information
SituationWhat is knownProposed step
The enquiry form does not deliver messagesThe team has reproduced the fault and confirmed where delivery stops.Act: prepare a correction and verify complete receipt.
A component has a security alertThe version and affected conditions are unconfirmed.Investigate with an agreed priority reflecting possible exposure.
Catalogue updates require manual workThe documented need depends on a data integration.Plan a review of the integration and alternatives.
The business is absent from an AI answerOnly one conversation exists, without recorded context.Observe under a protocol before attributing a technical problem.

This does not establish an order for every company. Here it explains why the form has a defined action while the alert needs checking. If the security flaw is confirmed to affect the website, review its priority with the responsible team.

Choose further reading around your question

When the concern is speed

Start by locating affected pages and tasks, then interpret the measurements. On WordPress, prepare a function and dependency inventory before retiring components.

When you need to understand security

Our vulnerability report guide separates observations and verification. For a client request, use the responsibility and evidence matrix.

When visibility changes

For falling traffic, review search changes and measurement. For generated answers, use a dedicated AI visibility record.

When considering an investment

Compare improvement, partial renewal and rebuilding. For a specific commission, agree scope and verification in the quote.

Give each priority an owner and an outcome check

Choose a next step you can justify, agree who will do it and identify the evidence needed to review it. Keep the remaining tasks in context; they do not all need to become immediate projects.

If a combined view is missing, a website assessment can connect security, performance and search visibility within agreed scope. Nodo91 can help you understand what needs attention and where to start. Implementation is scoped and quoted separately.