Winston's Crib

Five years of research, turned into a room planner

I made this from a problem I've been trying to solve for five years. I have a cat called Winston who is impossible to read, hates almost everything designed for cats and tells me he's bored by waking me up at 3am. I couldn't ask him what he wanted so I spent five years watching, guessing and iterating until I worked it out.

The builder is made from all of it. Everything in it comes from what Winston actually responds to. I built it myself using AI, which I'll come back to, because the interesting part isn't that I used AI. It's that I wanted to.

But first, the actual design problem.

A user who can't tell you anything

I got a cat who turned out to be nothing like I expected. He doesn't want to be held or pet and isn't interested in most things designed for cats. If he doesn't get what he needs during the day he will absolutely make that your problem at 3am.

The thing is I couldn't ask him what he needed. I could only watch, try things, get it wrong and try again. That's what this is about. Not just Winston but what happens when you can't ask your user anything at all.

Five years, at a glance

Winston. Age 5. Rescue cat. Extremely opinionated.

Winston likes being around people but don't try and touch him. He's curious about everything and plays only with specific toys in a specific way. He eats only dry chicken food and has strong feelings about which chair he sits in on any given day. He's not difficult, he just knows what he wants and the problem is he can't tell you what that is.

Some things never change while other things shift constantly depending on season, temperature or something I'll never fully identify. At 4am he becomes a completely different cat. Suddenly affectionate and happy to be near you but by morning he's back to normal. You learn pretty quickly to stop trying to make that make sense and just design around it.

After five years there are still open questions. Would he be happier with a companion? Is there a toy I haven't found yet? Would outdoor access solve the boredom or just move it somewhere else? Some questions stay open and that's not a gap in the research. That's just the reality of designing for a real user.

No formal process, just five years of paying attention

There was no formal process here. Winston lives with me and observation just meant paying attention over a long period of time. What that produced was this board. Everything I tried, everything that failed, everything that worked and the things he responded to that nobody planned for. That last category turned out to be the most useful of all.

From research to design

The research board started to show patterns I could actually work with. Winston liked climbing, watching birds, hiding and chasing things. He responded to things that felt unpredictable and real. He ignored almost everything designed and marketed for cats. With that I had enough to start making decisions.

What the research actually showed

Vertical space matters. Winston consistently sought height across every environment he's lived in. Doors, shelves, the tops of things. This never changed.

Realistic unpredictable movement beats mechanical toys. Anything with a fixed or repetitive pattern lost his interest fast. The toys that worked felt alive in a way the electronic ones never did.

Hiding is part of how he plays. Winston doesn't chase from an open space. He needs somewhere to hide and observe before he engages. Remove that and the play doesn't work.

Environmental changes worked better than products. Almost nothing marketed for cats made the green column. The things that worked were changes to his surroundings not things bought for him.

Novelty fades and needs refreshing. Even the things that worked stopped working if they were always available in the same way. Rotation and variation mattered as much as the thing itself.

Three things I built, and what each one got wrong

The cat wall

Winston climbed doors and always wanted to be higher up. The smaller cat towers I'd tried before stopped working once he got bigger and couldn't sit on them comfortably. A cat wall felt like the right solution. It gave him height, somewhere to survey the room and connected to everything I knew he liked about climbing.

The first version didn't work. The shelves were too small for him to settle on comfortably. He'd get up there and have nowhere to land so he'd come straight back down. He needed a reason to stay up there and without a comfortable place to settle he didn't have one.

I added a larger perch sized properly for him that connected into the existing wall. Then I trained him onto it using treats, essentially onboarding him onto something he had no reason to use yet. It worked. He started using it more and more and it became a reliable part of how he spends his time.

Even when you build the right thing you still have to bring your user with you.

The teepee

The logic was genuinely sound. The red column had already told me what didn't work. Soft beds, hammocks, tunnel toys, the watermelon bed. Everything cosy and enclosed had failed. But the purple column showed something different. Winston sleeping in specific chairs, wrapping himself in hanging sheets, gravitating toward new bed sheets. There was a clear pattern around fabric and enclosed spaces and I had two separate research threads pointing at the same conclusion.

The teepee wasn't soft or cosy. It was firm, it was on the floor and it gave him somewhere to go that wasn't a bed. It made complete sense. I bought it, washed it, put catnip on it, tried to make it feel familiar. He used it once.

What I'd missed was that the sheets weren't about enclosure or fabric at all. Winston liked them because of the coolness of clean cotton against his body. That's not something the research could have told me. I had a well reasoned hypothesis built on real observations and it was still wrong. The only way to find that out was to build the wrong thing.

Any product team that has spent weeks on a feature and watched users ignore it will recognise this feeling exactly.

The pillow fort

The bunny castle came from watching Winston sit in cardboard boxes. He liked cardboard, he liked enclosed spaces and the bunny castle gave him both. It was too small and he couldn't really get comfortable in it so I iterated. I got a bigger cardboard mansion thinking scale was the problem.

It wasn't the scale. Watching him I realised he wasn't interested in being inside the structure at all. What he actually wanted was to hide behind something and hunt. He wanted to feel like a predator with prey. That's a completely different need and no amount of cardboard was going to solve it.

That's when I stopped buying things and started thinking about what I already had. Big floor cushions, the one feather toy he actually likes, played in the specific way he responds to. Winston hides behind the pillows, I move the toy the way he likes and he plays properly. It's the most low-tech solution of everything I tried and it works better than anything I bought.

It also doesn't live on the floor of my apartment permanently which after the cardboard mansion was a genuine relief.

Not a solution, an environment

Winston still wakes me up sometimes. He's still unpredictable, still opinionated and still capable of ignoring something I was convinced he'd love. That hasn't changed and I've stopped expecting it to.

What did change was everything around it. The whining reduced. The 3am wake ups became less frequent. More than anything he just seemed to actually enjoy himself, which sounds simple but was the whole point from the beginning.

What I'd built without fully realising it wasn't a solution, it was an environment. The cat wall, the pillow fort, the bird feeder on the window, the one toy played in the one way he likes. None of those things work in isolation. Together they give Winston enough variety, enough stimulation and enough of a sense of having his own space that the boredom problem became manageable.

Some of the things that worked most weren't designed for cats at all. A pile of floor cushions. A bird feeder on a window ledge. Clean sheets. The things that came from just watching him rather than trying to solve him.

That's the part that stuck with me most. Not every user need has a product on the market that answers it. Sometimes the right solution is sitting in your observations the whole time and you just have to be willing to act on what you actually saw rather than what you assumed.

Cats are unpredictable and weird. So are users. The job isn't to find one perfect solution and ship it. The job is to keep paying attention.

Then I built it into a game

Once I understood Winston well enough, I turned the whole thing into a house builder. Everything in it comes from the research. The furniture, the surfaces, the things the cats respond to. It's the case study made playable.

I built it using AI, and this is the part I actually care about. Not that I used AI, but that I chose to. My role has been moving further into product and engineering, and AI is pulling design, code and strategy closer together than they've ever been. I'd rather run at that than wait to be pushed into it. So I picked a problem that had nothing to do with work, something I could think about purely logically, and used it to teach myself to build.

Superhero cats???

This was the first time I've hit engineering problems as the person actually trying to solve them, rather than the person handing them to someone else.

The cats kept jumping off the walls and I couldn't get them to move around the furniture naturally. I tried for a while and couldn't get it working. So I made them superheroes instead. Now they fly, and the capes make the flying make sense.

It's a silly fix and I'm not pretending otherwise. But the reason it stuck with me is what sat underneath it. AI gets handed to you as though it removes every obstacle, and a lot of the time it genuinely does. Then you hit something it can't just solve for you and you realise the hard part is still the hard part. The tool changes how fast you move. It doesn't make the difficult problems stop being difficult.

Here I could take the easy way out, because nobody is relying on these cats to walk properly. In real product work you usually can't. You have to sit with the problem and actually solve it. Getting to feel that difficulty directly, rather than watching an engineer absorb it for me, has done more for how I work with engineering teams than anything else on this project.

Also the cats look great in capes. 🦸