Interface design – start with the usage!

Years ago I worked at a tech company called Armature. What they did wasn’t very sexy – logistics software for supermarket chains that automated ordering and stock management. Lots of stuff that involves billions of boring pounds and will never have a Facebook group. By some curious quirk of fate, I found myself as head of usability testing and GUI design there for a short time before the company tanked spectacularly* (basically by betting £13 million of Schroder Ventures’ money on porting their entire product suite to Java. Duh.) 

One of the first projects I worked on was a product called Item Organizer. It was basically an interface to let a product or store manager look at an any item the company sold from a variety of different perspectives. It was pretty breathtaking. If you wanted to know what proportion of the weight of a tin of beans was taken up by the tin, you could look it up. You could see what categories of product it fell under and how it related to a product group as a whole, what stores it was held in, how much was in stock, the lead times, ingredients and so on to the nth degree. I can’t even remember the details but there was a shitload of information under the hood.

My job as GUI designer was to come up with a front-end to let people access all this information. I came up with a ‘tree’ menu. So you might click an option called Dimensions and be presented with sub-items for weight, height, depth etc. And sometimes these levels would go four deep. We spent a lot of time arguing how to indicate which branches of the tree were open and editable, eventually settling on some kind of arcane traffic light system with beautiful little icons that looked like glass balls of various colours. Purely because I’d done a Photoshop tutorial on how to make glass balls, if I’m being honest.

Anyway: it made for a hell of a demo. The sales guys were as happy as shit, because if it’s one thing that impresses the kind of people who sign P.O.s for £800,000 pieces of software it’s a demo that looks beautiful and illustrates how much data it can hold.

But then, I got a snippy email from one of the consultants who were actually implementing the thing down in the bowels of some supermarket somewhere, bitching because they couldn’t understand how they were supposed to use it. The caused ructions so severe that I found myself sitting alongside a couple of Product Managers composing a 15,000 word email that was just a big bunch of defensive politics, along with a hidden but pointed suggestion that they try reading the help files. Dumbclucks.

For a fraction of a second after we hit the ‘send’ button, I felt the warm glow of satisfaction at having protected my turf from amaetuer interlopers. I was the GUI designer and I’d put in the hours, and those fucktards had better appreciate what a bang-up job I’d done. But then, in the quiet recesses of my soul, I harboured some little doubts. Eventually, I called the guy who’d sent the email to find out what his gripe actually was:

Me: Hi Implementation Bloke, it’s me – the guy who designed that GUI you were bitching about? What exactly’s wrong with it?

Implementation Bloke: Do you know who uses this software in real life?

Me: Err… store managers who want to look at an item from the store against various metrics?

IB: No. Data inputters.

Me: Wait a minute. You mean those $4 an hour guys with a printed spreadsheet and a ruler?

IB: Yep. And that stupid interface of yours means that they have to stop after pretty much every column to look up at the screen and navigate to the right place. It takes whole minutes to put in 6 bits of information, which is mainly all there is.

Me: Oh.

So there we had it. The premise of the software was all to cock from the start. Our imagined user was the guys we were demoing the software to – basically some managers from Shitkicker Supermart inc. whose involvement ended when a cheque got signed. Meanwhile, the poor schlubs actually using the thing would be reduced to a state of nervous anxiety.

Flushed with the zealotry of youth, I marched up to see various bits of internal management brandishing bits of paper demanding to know why we’d got it wrong. But they were all earning £80k per year and I was just a bloke who did things with pixels so my opinion counted for naught.

I even hit on the perfect solution – a lightweight web interface where the front end could be reconfigured to actually match the order of the columns on the printed sheet. A trivial exercise that would have boosted usability about 6 million %. Anyway, it didn’t matter because the company went bust.

Actually – reverse that. The company went bust because that did matter. Building for demos was a mug’s game then – and every time you see an ecommerce site with an innovative interface, it’s still a mug’s game. I’m out of the design game now, but the lesson has stayed with me ever since. If you’re designing for a purpose – make sure you know what that fucking purpose actually is before even you start.

—————————————————————– 

* I’m pleased to report that the top team there dusted themselves down and secured funding for another crack at this stuff. Let’s hope they’re reading!