I have spent thirty years as a designer, so what follows is not an attack on design. It is the opposite. I care about it enough to think most websites get built in the wrong order.
Here is the usual order.
Somebody decides they need a new website. There are conversations about how it should look. Moodboards, colours, a homepage concept, some pages sketched out. It gets built, everybody is pleased with it, and then somebody says “what about SEO?”
At which point a person like me is handed a finished website and asked to make it findable.
That is backwards, and it is why so much SEO work feels like patching.
What actually goes wrong
When search is considered last, you inherit a structure that was decided for other reasons.
The pages that exist are the pages somebody thought of during a design conversation. Usually: home, about, services, contact, maybe a gallery. Perfectly sensible as a brochure. But the questions your customers actually ask do not map onto those five pages, so there is nowhere for the answers to live.
The navigation was designed to look clean, which usually means fewer items, which usually means the specific pages you needed got merged into a general one.
The homepage says something atmospheric, because that is what looks good in a mockup, rather than saying plainly what you do and where.
And the individual services are all on one page together, because that laid out nicely, which means none of them can be the answer to anything specific.
None of those are design failures. Every one of them is a perfectly reasonable design decision made without knowing what the site had to achieve.
Then I turn up and can only really tinker. I can rewrite the titles, tidy the headings, add some copy. What I cannot do, without starting again, is create the pages that should have existed from the beginning.
The order I would use instead
Work out what the site has to be found for. Then let that decide the structure. Then design it.
Before any of that: where is this business going?
This is the step almost everybody skips, and it is the one that decides whether the site is still fit for purpose in two years.
Most websites are built for the business as it is on the day of the brief. Which sounds sensible until you consider that the business is not going to stay that way. You will add a service. You will start covering another town. You will decide the work you actually want is not the work you currently get most of.
So the first question is not what do you do. It is where do you want to be in twelve months, and what will you need to have published to get there?
If somebody wants to move from general building work towards period property restoration, that changes everything downstream. Different pages, different structure, different emphasis, different content plan. Build for what they do today and you have built something that fights them next year.
First: what questions must this site answer?
Not what pages do we want. What does somebody actually ask before they buy from you. What does it cost, do you cover my area, how long does it take, can you do the specific thing I need, what happens if it goes wrong, have you done this before.
Get those written down and you have most of your site already.
Second: what pages does that require?
Now the structure builds itself. Each significant service gets its own page, because each one answers a different question. Each area you cover gets treated properly. The questions that come up repeatedly get somewhere to live rather than being squeezed into a paragraph.
You end up with more pages than a designer would naturally propose, and each one has a reason to exist. That is the opposite of how most sites end up, where the page count comes from the layout.
Third: what does the navigation need to do?
Now it can be designed around what people are actually looking for, rather than around what fits neatly across the top.
And only then: what should it look like?
By this point the design brief is far better than it would have been, because you know what each page is for. Designing a page when you know its job is a completely different exercise from designing a page because the site needs an about section.
Foundations that can take the weight
Here is where this turns from a philosophy into something practical, and it is the bit I care most about.
A website is not finished when it launches. It is going to grow. More pages, more services, more towns, more questions answered, month after month. That growth is not an optional extra, it is how you actually get found.
So the structure has to be built to take it.
That means the navigation has to make sense with forty pages, not just with six. It means the URL structure has to accommodate services you have not thought of yet. It means the internal linking has to be something that keeps working as things are added, rather than a one-off arrangement that gets messier with every new page. It means the design has to hold up when there are three services in a section rather than one.
None of that is visible on launch day. All of it decides whether year two is easy or a rebuild.
And this is precisely why I work monthly rather than selling a website and disappearing.
Not because a subscription is a nicer business model, though I will not pretend it is not. Because a site built properly is designed to grow, and something has to actually do the growing.
A one-off build is a snapshot. It is the business as it was on the day it was signed off, frozen. Eighteen months later it is either out of date, or somebody has been bolting bits onto a structure that was never meant to carry them, which is how sites end up slow, confusing and impossible to fix without starting again.
The plan is the product, in other words. The website is what the plan gets built on.
Why this makes the design better, not worse
I want to be clear about this, because “SEO first” sounds like a designer being told to stop having ideas.
It is the opposite. Constraints improve design. A designer given “make a nice homepage” produces something decorative. A designer told “this page has to convince a homeowner in Poulton that we can fix their roof, and get them to ring within five seconds” produces something that works, and usually something better looking too, because it has a point.
The worst websites I have seen were not badly designed. They were beautifully designed and pointless. Somebody spent a great deal of care on something that was never going to bring in any work, because nobody established what it was for.
That is the actual waste. Not ugly websites. Lovely websites that do nothing.
What this looks like in practice
For me it means the search thinking happens during the conversations, not afterwards.
When I sit down with somebody at the start, we are not talking about colours. We are working out what they do, who for, what makes them different, what their customers worry about, what they get asked constantly, which towns matter and why, and crucially where they want the business to be a year from now.
That conversation produces the site map. The site map is really a list of questions with pages attached to them. And the design comes afterwards, at which point it is a genuinely enjoyable job, because I know exactly what every page has to do.
The internal linking, the structure, the speed, the headings, none of that is a task list I run through at the end. It is the natural result of having decided what the thing is for. Nobody bills me for it separately because it is not separable.
This is also why I am cheerful about giving so much away publicly. The advice is not the hard part. Knowing what a specific business needs, and building in that order, is the hard part.
Which is why bolting it on afterwards costs more
A practical warning if you are about to commission a website.
If somebody designs it now and thinks about search later, you will pay twice. Once for the build, then again when it turns out the structure cannot support what you needed, and the honest fix is a rebuild rather than a patch.
That is not the designer being incompetent. They were never told what the site had to achieve, so they optimised for the only brief they had, which was that it should look good.
Ask, before anybody opens a design tool: what does this site need to be found for, and what pages does that require?
If the answer is that you will work that out later, you have just booked yourself a second project.
The bit I got wrong for years
I did not always think this, and I should be straight about how I know.
I co-owned a design agency called Freebird. It is not trading any more. We built websites the way everybody did then: start in Photoshop, get the design right, make it look excellent, then build it.
And they did look excellent. I am still proud of a lot of that work.
They did not rank particularly well.
We got away with it for a good while, because at the time you could. A handful of footer links and a bit of care was frequently enough to get a small business showing up. Search was a considerably softer game and design agencies could ignore it without paying much of a price.
I did not know enough about it to see what was coming. I came up through print, where the job genuinely was to make the thing look right, and the idea that structure should be decided before appearance would have sounded backwards to me.
So this is not a lesson I read somewhere. It is one I got wrong for years, in my own business, with my own clients, and it took a long time to properly admit.
INKON exists largely because of it. Building sites where the search thinking comes first, and the design serves it rather than the other way round, is the specific thing I wanted to fix, and it is why I spent months building the way I work now rather than just carrying on.
Peter, who taught me most of what I know about this, used to say: make it look beautiful, and make the offer visible, so it turns into new business. I have quoted that a lot.
It took me an embarrassingly long time to notice the order he put it in was not accidental, and that the second half is the bit that decides whether the first half matters.
Want a website built in the right order? Get your free audit, or email hello@inkon.co.uk. I read every message myself.