How we work

1. A conversation about the problem

Not about the number of pages. What we want to establish is what the site is for: who is meant to arrive, what you want them to do, and what currently stops that happening. A site that gets ten enquiries a month from the wrong people has a different problem from one that gets none.

If what you need is smaller than a new website, we will say so. Sometimes the answer is a faster host, or fixing what is already there, or writing three pages that do not exist yet.

2. A structure, agreed before anything is designed

A list of pages, what each one is for, and what sits at which address. This is dull and it is the single most useful hour of the project, because every later decision depends on it and changing it afterwards is what makes projects expensive.

For a site that is replacing an existing one, this step also produces the redirect map: every URL that exists now, and where it will point afterwards.

3. Design, on real content

We design with your actual words in place, not with placeholder text. Placeholder text hides every real problem: the heading that is four words on the design and nineteen in reality, the paragraph that is three times longer than the space allowed for it.

That has a consequence worth being upfront about, which is step four.

4. The content, which is where projects stall

We can write it, you can write it, or we can work from what you already have and edit. What does not work is waiting for it, so we would rather agree at the start which pages you are writing and which we are, with dates against them.

5. Build, then check it properly

The build itself is the predictable part. What we check before launch:

• Every page loads, including the ones below the top level, which is where a permissions problem shows up as a "forbidden" error while the homepage looks fine
• Every internal link resolves, so there are no links to pages that were renamed
• The site works at phone width, tested at phone width rather than by shrinking a desktop window
• The certificate is not only issued but actually bound to the site, which is a distinction that has caught us out before and is worth checking rather than assuming
• Redirects go where the map says, including the ones that should say "permanently gone" rather than "not found"
• A real sitemap exists, lists every page, and carries honest dates rather than today's date on everything

6. Launch, and then the boring weeks

Going live is a DNS change and is usually the least eventful part of the day. The weeks afterwards are where the useful information is: what pages people actually land on, which ones they leave immediately, and what search engines do with the site.

We would rather look at that with you after a month than hand over a site and disappear, because the first month is when the cheap improvements are obvious.

What we need from you

• Someone who can make a decision. One person, not a committee reporting back
• Access to the current site and its domain, or the details of whoever holds them
• The content you said you would write, on the date you said
• An honest answer about what has already been tried, including anything that went badly. It saves us both from repeating it