Mistakes I made designing a coworking space I'd never actually worked in

When I first opened my laptop at a coworking space in Lisbon, I realised just how terrible the project I'd designed two years earlier really was. The booking system there was poor, but at least it had been built on an understanding of how people actually settle into a seat, how they go to get coffee, and how they avoid phone calls. What I'd built sitting in my studio in Brighton was not.

Around 2022, I took on the UX for a coworking space booking platform at a real estate startup in London. The client was enthusiastic and the budget was decent. The problem was me. At the time, I wasn't a digital nomad. I worked from home, and the occasional trip to a café in Brighton was the entirety of my "remote work experience." I'd only ever seen coworking spaces in photos. Bright, clean photos, the kind with a MacBook sitting on IKEA furniture.

Noise didn't show up in the data

My first mistake was being completely oblivious to noise. I divided seat types into two categories: "open desk" and "private room." I thought it was clean and logical. The heatmap data showed high open desk usage, so I concluded that most people wanted open desks.

It wasn't until I spent a week at Punspace in Chiang Mai that I learned the truth. A significant number of the people using open desks actually wanted a private room but were sitting at open desks because of the price, or because private rooms were fully booked. And the type of noise was the issue. Keyboard sounds are fine. Zoom call sounds are the problem. Once you've experienced someone screen-sharing and repeating "Can you hear me? Can you hear me?" echoing across the entire open desk area, you understand in your bones why a "quiet zone" filter is necessary.

If I were redesigning it, I would categorise spaces not by seat type but by noise tolerance level. "Calls allowed," "quiet work," "complete silence." It's a taxonomy you simply cannot come up with until you've actually sat in a coworking space.

Outlet placement is UX

When I was drawing wireframes in my Brighton studio, I didn't even consider marking "power outlet availability" on the seating map. A desk was a desk. I assumed every desk would have an outlet.

At a coworking space in Canggu, Bali, I moved seats three times. A seat with an outlet, a seat with an outlet that actually worked, a seat with a working outlet that wasn't directly under the air conditioning. Only about a third of all seats met these three conditions simultaneously. Those seats were all taken by 8 a.m.

The booking system I'd built had none of this information. Users would book "Desk A7," but until they arrived, they had no idea what A7 was actually like. Whether it was by the window, near an outlet, or next to the bathroom. Hotel booking sites show dozens of photos of a room, so why doesn't a coworking space booking system tell you the conditions of a single seat? Because I built it that way.

Nobody wants to stand at a check-in screen

For the London project, I carefully designed the check-in flow. Upon arrival, you scan a QR code at a tablet kiosk, a welcome message appears, today's Wi-Fi password is displayed, and a screen comes up asking you to agree to the terms of use. Five steps. I thought it was a clean onboarding experience.

What I actually saw at a coworking space in Medellín was this: people walk in the door and sit at any empty seat. They put their bag down, open their laptop, and ask for the Wi-Fi password. That's it. Nobody stood in front of a kiosk going through five steps. Most places didn't even have a kiosk. For someone who's just arrived in an unfamiliar city, caffeine-deprived, dragging their luggage in, a five-step onboarding is practically an insult.

If I were doing it now, I'd make check-in happen automatically. Whether through geofencing or Bluetooth via the app. A good check-in is one where the user doesn't even realise they've "checked in." The Wi-Fi password should always be visible at the top of the app's home screen. Terms of use agreement should happen once, at initial sign-up.

The illusion of monthly billing

When designing the pricing structure, I created a "daily pass / weekly pass / monthly membership" model. The classic three-tier structure you see on SaaS pricing pages. I thought that emphasising how the monthly membership was the most economical option would increase the conversion rate.

After spending a year as a digital nomad, I realised that staying in one city for a full month is a luxury in itself. Far more often, you stay for two weeks and move on. Even when you do stay in the same city for a month, you don't go to the same coworking space every day. Tuesday at a café, Wednesday at the accommodation, Thursday at the coworking space. If you buy a monthly membership, you actually end up using it for roughly ten days out of the month. The effective daily cost of a £200 membership works out to £20. More expensive than a £15 daily pass.

If I were building it again, I'd propose a credit-based system. Rechargeable, with a generous expiration window, usable at partner spaces across multiple cities. There's a reason WeWork moved in this direction.

The common thread running through all my mistakes is simple: I designed for an abstract entity called "the user" while sitting in an office. I created persona documents, drew user flow maps, and analysed heatmaps, but I failed to understand that this user was a real person. Someone wandering around looking for an outlet, sitting in a corner by the bathroom to escape Zoom call noise, piling luggage to one side and feeling anxious about it. Now, as the first step of my research, I work in the actual space whenever possible. Three days minimum. Convincing clients that the cost of a plane ticket should be included in the research budget is half my job these days.