Equinix, growing UX maturity inside an enterprise

Equinix owns most of the physical facilities that internet traffic moves through. If they vanished tomorrow the internet would collapse. AWS, Google Cloud, Apple, Uber and Airbnb all run inside their data centres.

The business problem was not awareness, it was explanation. The website was written in internal language, organised like the company's org chart and almost impossible to navigate as a customer. On top of that, the company had no prior experience working with designers and no shared idea of what design was for.

I was hired as a senior designer but the role was wider than that from week one. I mentored the small group of in-house designers, defined product strategy alongside the PMs, ran user research, chose the team's tooling, and kept a single holistic vision across every new initiative.

Most of the job in the first months was not Figma, it was conversations. Explaining what design is, why we test with users, why we cannot just paste the print brochure onto a webpage.

Equinix's online presence was becoming critical for lead generation and customer education, but it was not doing the job. The writing was convoluted, the content hierarchy mirrored the company's internal structure and customers told us, in research, that they could not work out which product solved their problem.

On top of the content problem, there was a culture problem. There was no design thinking practice, no shared research, and no agreement that customer needs should drive the site. That made the work bigger than a redesign.

I split the work in two tracks. Track one was visible, customer-facing redesigns that would generate evidence. Track two was the quieter, slower work of building a design practice: hiring, mentoring, choosing tools, writing principles and changing how decisions were made.

Both tracks fed each other. Every shipped redesign gave me data I could use to argue for the practice. Every conversation about the practice gave me the air cover to ship the next redesign.

The first real ship was a self-serve product finder. Customers could search for available cloud services at a particular Equinix location instead of trying to decode the navigation.

It got the highest engagement of any launch in website history, executive leadership called it out as one of the best launches the team had done, and stakeholders who had been critical of the design process became the loudest advocates for it.

There was no quantitative research in place. I installed Hotjar, since Fullstory was outside the budget, and we started using Google Analytics seriously, not for vanity numbers, but to tell stories about behaviour we could put in front of stakeholders.

Heatmaps, session recordings and funnel drop-offs became the language that finally got senior leadership on board. "Look at this twenty-second clip" ended more arguments than any deck I could have built.

Six months in, the language in the room had shifted. Where stakeholders used to say "let's put this print brochure on the website", they started asking "is this good for the user?" and "what data do we have?".

That shift was the real product. It meant the next designer who joined would not have to fight the same battles I had.

The data centre product pages were the highest-stakes part of the site. They had to explain a deeply technical product to a mixed audience of IT directors, procurement leads and engineers.

I rebuilt the pages around the customer's question, what does this location give me, what does it cost me in effort to get there, who else is already running here. Side by side, the original and the redesign tell the story.

The redesign led with the answers to the three questions customers actually ask. Specs and technical detail moved into a clearly labelled section further down the page, so the people who needed depth could still get it without burying everyone else.

Once two pages were live in the new pattern, I started a holistic style guide and design system: typography, colour, spacing, button and form patterns, page templates and content rules.

The system was deliberately small to start with. The aim was not to ship a finished library, it was to give the team a single source of truth they could grow with as more pages were rebuilt.

The homepage was the hardest piece of work because every team in the company had an opinion about it. I went back to research, used the data we now had from the product finder and the data centre pages, and used that to defend a much simpler structure.

The new homepage cut the number of competing messages, led with the customer's job to be done, and gave the product finder a clear primary entry point. It set the visual direction the rest of the site would follow.

Most of the energy in an enterprise goes into alignment. I ran short, structured stakeholder reviews every two weeks, with three things on the wall: what we learnt from research, what we changed because of it, what we are going to test next. No long decks, no surprises.

With engineering, I kept the loop tight. Anything that was expensive to build got rescoped together, not over a wall, so we could ship a usable v1 now and grow into v1.1 later. The same tactic I had used on smaller teams worked just as well here, it just took more conversations.

The redesigns shipped, the data was strong enough to defend, and the leadership team approved budget and headcount to hire a permanent design team and rebuild the site from scratch. The grand vision was a design system the rest of the industry would notice.

Even after I left the company, the patterns and the practice stuck. Going back to the site years later, you can still see the bones of the work in how content is organised and how decisions get made.

Two lessons I carry into every senior role since. First, you cannot evangelise a practice in the abstract. You ship one visible thing, you measure it, you let the numbers do the arguing. Second, mentoring is the multiplier. Every hour I spent with the in-house designers came back as ten hours of momentum I did not have to create myself.

The honest version of this story is that not everything worked. Some redesigns were watered down by committee, some research never reached the people who needed to see it. The proof of the process was that the things that did land kept landing after I was gone.

If I picked this up again I would invest earlier in research operations, so insights got reused instead of rebuilt each time. I would also formalise the design system governance from day one rather than letting it grow organically, because at enterprise scale informal does not survive a re-org.

Handover was less a single moment and more the thing I was building the whole time. Every redesign shipped with annotated Figma files mapped to the live site, every research finding lived in a shared workspace, and every pattern in the early design system had usage notes a new designer could pick up without me in the room.

When the permanent team came in, they had a system, a research baseline, a backlog and a written point of view on how design worked inside the company. That is what I am most proud of from the project, the practice stayed after I did.