The project process

A useful brief tells us what the first release needs to do. We agree on the audience, the main tasks, the constraints, and what we will check before launch.

From brief to release

I sketch the important screens and review them with you before building the components. During implementation I check the layout at different widths, test the main flows with a keyboard, and exercise API failures where relevant. At launch I check the deployed pages and document how to update them. We can agree on ongoing support as part of the scope.

Begin with an example everyone understands

Rather than asking you to specify an entire website in the first conversation, I ask you to describe a realistic visit. Who arrives, what brought them there, and what would a successful visit look like? A prospective customer comparing two services has different needs from a staff member processing the tenth request of the morning.

We use that example to identify the essential pages, the information each one needs, and any work happening outside the site. This is also the time to mention an existing brand guide, a system that must keep working, a date tied to an event, or the person who will approve the content. A constraint raised early can help us make a decision. The same constraint discovered shortly before launch can undo one.

Settle the first release

Once we understand the main task, we can decide what belongs in the first release. I like to state that in language we can test. A visitor should be able to choose a service, describe a project, and review a message with the right contact details. We can walk through that statement and agree when it is true.

Other ideas can still have merit. If a new request changes the amount of work or introduces a dependency, we can look at its effect on the release together. Sometimes it deserves a place in the current scope. Sometimes the existing flow needs to be in people’s hands before we know how an additional feature should work. The point is to make that choice visible while it is still a choice.

Review the design with real material

I sketch the important screens and place representative content into them. A real service description tells us how much space a heading needs. An actual field label can reveal that the mobile form is cramped. A sample record can show which pieces of information people need together when they make a decision.

During review, I ask you to try a task. Find the right service, decide whether it fits, and take the next step. Explain what you expect a control to do before you use it. Your answer can uncover a misleading label or a missing piece of context. This kind of feedback is easier to act on than a general request to make a design feel more intuitive.

Build the complete path

I connect the important parts early enough for us to try the flow as a whole. If a dashboard needs data, we agree on sample responses. If a form hands off to another application, we check the message at the destination. A series of attractive individual screens can still have a missing step between them.

The build includes the conditions people will encounter after the first demonstration. Text may wrap differently on a phone. A visitor may go back to correct an answer. A request may take longer than expected. When an issue appears, I trace it to the component, layout, or data rule responsible. That gives the fix a better chance of holding up elsewhere on the site.

Test before inviting visitors in

Before release, I check representative screen sizes, follow navigation links, and exercise the main actions. I use a fresh visit as well as a return visit, because saved preferences or previously completed steps can change what a person sees. For a site with a contact flow, I check the information that reaches the final handoff as well as the form people fill in.

I also look at the pages around the main journey. A missing address should lead to a useful error page. A footer link should take someone to the service its label promises. A long page should leave its final action reachable on a small screen. These details matter most to people who arrive somewhere other than the homepage.

When there are several reviewers, a shared example keeps feedback manageable. Include the page address, what you tried, what you expected, and what you saw. “The service choice disappears when I go back from step two” gives me an action I can reproduce and a clear way to confirm a fix. For a visual issue, include the approximate screen size and whether you had enlarged the text. Once an issue is fixed, the person who first reported it can repeat the same sequence. That closes the loop more usefully than a general statement that the page has been updated.

Launch with a clear handover

A deployment deserves its own check. I open the published pages, follow their links, and verify the actions included in scope. The working preview and the public site can differ in their addresses, configuration, or connections to other services. A quick pass through the real journey catches issues that a local preview cannot settle.

For handover, we look at the updates you are most likely to make. That could mean editing a service description, changing an enquiry destination, or adding a project to a shared template. We agree on who owns those updates and which changes need development work. If ongoing help is useful, we discuss its scope separately so you know what to expect after launch.

You can start with a rough brief. The most useful things to bring are one example of the task, the audience who needs it, and any commitment that will shape the first release.

Essential storage is always on. It remembers your choices and supports basic site features.