BizDigital: an agency site redesign that started as an unsolicited pitch

How it started

I was studying for a diploma in Brunei and needed an internship placement. Going through the agencies on the list, BizDigital was the one doing work that was actually designed rather than assembled, so it was the one I wanted.

Applying with a CV would have put me in the same pile as everyone else on the same course. Instead I redesigned their homepage and sent them that. It was not a critique and it did not tell them their site was bad. It was the site they already had, restructured: the same four services, the same claims, the same clients, put on a ground that held them together and in an order somebody could follow to the end.

They gave me the placement, and once I started they asked me to take the concept across the rest of the site. That is the reason the design took under a month. The page that usually takes longest to agree had been agreed before my first day.

The problem

An agency's own website is the one piece of work every prospective client sees, and this one was arguing against the agency. It read as a template that had been filled in rather than a site that had been designed, which is the worst thing a design agency's site can look like. Three things were doing most of the damage.

White page, black text, nothing in between

The page had two tones and no steps between them. With no midtones there was nothing to separate one section from the next, so headings, body copy and links all carried the same weight and the eye had nowhere to land. A visitor scrolling it could not tell where they were.

The images were doing the work, and often did not arrive

Layout depended on large images with very tall aspect ratios. When they loaded slowly, or failed, the sections they belonged to collapsed into empty bands of white. The page managed to feel cluttered and empty at the same time, and on a slow connection it looked broken.

No section ended in anything to do

The content ran as one long column with no clear breaks and no call to action attached to any of it. Nothing asked for the enquiry, and the contact form sat alone at the bottom of a page most people never reached.

What I did

Built the site outward from the page that was already agreed

The homepage concept set the ground, the card, the type scale and the order of a page. Every other page was then a question of which blocks that page needed, not a fresh design. Working outward from an agreed page is what let ten pages be designed in under a month by one person.

Put the page on a gradient and lifted the content onto cards

The gradient runs deep purple at the top of every page to a pale lilac at the foot, which gives the page a direction and tells a reader roughly how far down it they are. Content sits on light cards raised off that ground. Cards solved the problem the old site had with images: a card has its own edges and its own fill, so a section still looks like a section when the picture inside it has not arrived.

It also fixed the hierarchy. A card is a boundary, and once every block of content had one, a visitor could see where a section started and finished without reading a word of it.

Gave the five service pages one template

Design, development, marketing, media and the overview page all say the same kind of thing about different work, so they run the same skeleton: navigation, a quote stating the promise, a two column grid of service cards, one call to action and the footer. Only the content changes.

A shared template was the practical decision as much as the design one. The developer built the pattern once and filled it five times, which is most of why the build took a month instead of three.

Designed against what the build could actually do

The site was going to be built on WordPress with Elementor, by the studio's developer and not by me. So before committing to anything that moved or stacked in a particular way, I asked him whether it was buildable in that stack and what it would cost in performance. Several ideas changed at that point, and changing them on an artboard is free.

This was the first time I designed something I was not going to build myself, and the habit came out of it: agree the constraint with the person who has to live inside it, before the design assumes it away.

Settled the language once and then stopped deciding

One ground, one card fill, one corner radius, one shadow, one accent. The accent is reserved for the thing to press and for the line art, so nothing else on a page competes with the button. Fixing this early is what made the remaining pages fast, and it is why ten pages designed over four weeks still look like one site.

The senior designer left the design to me and reviewed it as it went. His input landed on the things I had least experience with at the time: whether an animation was worth its cost, whether a purple was still legible behind white text, and how to hold the whole thing together with styles in Figma instead of by hand. The design was mine; the reasons several details survive contact with a real build were his.

What happened

The site went live and stayed live. I was there for four months of it, which was long enough to ask my manager whether it had actually changed anything. He pulled up the agency's analytics and read the before and after off it: engagement up around 30%, lead generation up around 20%.

Those are the agency's figures, not mine. I did not have access to the dashboard, I did not record the base numbers, and I cannot reproduce them now, so they are worth exactly what a client's own reporting is worth. The direction is the part I would stand behind: the site started producing enquiries it had not been producing before.

Where it stands now

bizdigital.biz today is a later redesign by someone else, so there is nothing live to link to. What is on this page comes from the original artboards, with one archived capture of the old site for the before. That is also why the work is dated plainly to 2023: it is a record of what I could do in my second year, not a sample of how I design now.

The agency offered me a job at the end of the internship. I took a higher paying role with a Singapore agency instead, and left on good terms.

What I took from it

The pitch is the part I have kept. Sending finished work to somebody who had not asked for it did more in one email than a term of applications, because it answered the only question they had. I have opened every serious conversation the same way since.

The other lesson was about handing work over. A design that cannot be built in the stack it is destined for is still only a proposal, and that difference shows up when it is too late to change cheaply. Asking the developer first cost me a few conversations and saved the project a month.