I hate bad service design

Three examples from the last few months, three different industries, same exact mistake.

The indoor playground

I took my toddler to an indoor play park. Under two, entry is free, but only if you register online first. At the till, I got asked to leave the queue, dig my phone out, and fill in a web form before they'd let us in. Skip that and it's 14 euros, for a kid whose entry is supposed to be free.

The fix is embarrassingly simple. One QR code on the counter. One sentence: "scan this, fill in your info, you're in for free." Ten seconds, no leaving the line, no fumbling with a phone while wrangling a pushchair mid-transaction.

Instead the whole thing was built so the customer has to physically step out of the service moment to go solve the company's own internal process for them. That's not a small bit of friction. It's a queue-stopping mess, because the family behind you now eats the cost of you filling in their form for them.

The healthcare waiting screens

Helsinki's public healthcare waiting rooms now run fully automated queue systems. Good idea on paper. The execution is a joke. The screens update on their own schedule, and what they show isn't remotely useful, no sense of how many people are ahead of you, no real estimate of how long you'll be sat there staring at a number that means nothing.

This is a trivial fix. Take the average handling time for that time of day, run the maths, show a wait estimate next to every number in the queue. You don't need AI, you don't need a new platform, you just need someone in the building who thought about this from the patient's side of the counter instead of the system architect's.

Automation isn't the villain here either. The villain is building automation without anyone asking what a person actually needs to know while they're waiting. A number on a screen with zero context isn't information. It's just proof you've been filed away and forgotten.

I'm by no means the smartest person in the room, but if I struggled to work out what I was meant to press and how the whole thing was supposed to function, a lot of people are going to be stumped as fuck. And these aren't optional systems you can just avoid if they're not for you. This is a doctor's waiting room. The people using it skew older, skew sicker, skew less patient with a screen that won't tell them anything useful. Building a confusing automated system for that exact audience isn't a rounding error. It's the whole design brief, missed.

Finnair's automated bag drop

Finnair's new automated baggage drop points are the same failure wearing a different uniform. The pitch: cut check-in staff by having passengers handle their own bags. The reality: the automated points need more staff standing around explaining how to use the machines than the check-in desks ever needed behind the counter.

This is exactly where "efficiency" stops being efficiency and turns into cost-shifting with a straight face. The company saves on desk wages and then has to pay people to stand next to the robots, coaching confused travellers through them one at a time. From the passenger's side, the whole process is slower, more confusing, and more stressful than the desk it replaced. On a spreadsheet, the investment probably looks great, because a spreadsheet doesn't have a line item for the guy standing baffled in front of a machine while his gate closes.

Same mistake, every time

In all three cases, automation itself isn't what went wrong. Automation is a tool, not a virtue you get to claim just for using it. What went wrong is that someone optimised for one number, staff costs, throughput, whatever it was, instead of optimising for what the person on the other end of it actually experiences.

Good service design doesn't mean doing everything by hand. It means that when you do automate something, somebody in the room actually answers the question: is this easier for the person who has to use it, or just easier for the person who built it and gets to go home?

If the answer is the second one, you haven't created efficiency. You've just relocated the friction somewhere nobody's bothering to measure it anymore.