I disagree with half the of the assertions in this article. I don't like how the coke freestyle machines work. They touch screens are finicky, the soda doesn't taste as good, and the interaction feels a little over-the-top. I think the simplicity of most gas pump interfaces is actually a feature. My main gripe is that they are slow. While I don't think gas pump ux is amazing, it seems like the author is conflating visual design with effective user experience.
> They touch screens are finicky, the soda doesn't taste as good, and the interaction feels a little over-the-top.
Those aren't things in question here
> it seems like the author is conflating visual design with effective user experience
often the fact that people separate those things so much is an issue in itself. visual design is definitely a big part of an effective user experience and a user's feelings about the product they are currently using
much like how the feeling of responsiveness of an interface can be manipulated by simply changing the duration of a few animations (but not the actual duration of the action)
Sure, that's fair. I said visual design, but I think I might have meant more simply "look". Design implies some intentionality, but it felt like the commentary was focused simply on how it looked and not how that actually impacted the design. For example, the resolution of the pin graphic doesn't really impact the design significantly.
Coke Freestyle is a poor example of good UI. The system is so underpowered it can't reliably detect touch events or even the physical dispense button. It's running CE on something like an early 00's reference PDA platform that can't push the pixels to a big screen and handle all the embedded IO at one time. That's why the 2fps animation is such a joke.
It's funny to watch young kids using quick taps and get nowhere, expecting it to work like an iPad.
The Pepsi equivalent puts Freestyle to shame on UI polish.
One problem being (badly) addressed in the images, in my unskilled opinion, is that is it possible to finish the transaction without replacing the nozzle in the holster (and then, presumably, drive away with the nozzle still in the car). While no kind of UI expert, I do believe that good UI should guide the user into a correct flow of actions and make it difficult to make a mistake.
I have a hazy memory that in the field of ATM design, the reason you don't get your money before you take your card back is to ensure that people take their card back; because people were focused on getting money, they would leave the card behind as soon as the money was in their hand. The solution was to make getting the card back a necessary step towards getting the money.
The analogy doesn't quite hold; people want the fuel, at the point they have the fuel (and are effectively done getting what they want) the nozzle is still in the car, but is there a way to make finishing the transaction more difficult without replacing the nozzle in the holster? Perhaps if paying by card, the machine could hold onto the card until the nozzle is replaced, or in some other way make payment impossible without replacing the nozzle. That would work where I am, but for all I know there are systems out there in which people complete payment in advance, get their card back, and are then dispensed a pre-agreed amount of fuel.
They could have a mechanical version of this, although the placement would have to be figured out, cars have different dimensions and park differently at the pump.
A rental I had required you to use the ignition key to open the inner fuel cap, and you left the key hanging from that cap as you refuel This makes sure you turn off the engine, and it also makes sure you won't leave without replacing the fuel cap (and obviously that can only be done without the pump nozzle in the way).
They used to have this in Formula 1, but still the driver would sometimes race off before the fuel nozzle was out. After a series of incidents, a couple nearly really serious, Formula 1 scrapped refuelling mid-race - the cars now start with enough fuel to finish the race.
I don't get why US gas pumps are so complicated at all. Here, I drive up to the pump, grab the right nozzle, fill up, go inside to the cash register and pay. There are literally bo buttons on the pumps. Fiddling with the pump UI in the US has always been a slower process for me.
> Here, I drive up to the pump, grab the right nozzle, fill up, go inside to the cash register and pay.
That was exactly how it worked in the US until about 20 years ago. But then stations started to require prepayment because there were a number of people skipping that important final step.
Well you're talking about UX and the article is mostly about UI (touches on the experience here and there). Arguing that the pictures themselves look amateurish.
In Europe most pumps are a lot simpler and the only screen is the segment LCD giving you the price and liters. Sometimes you'll have a couple of buttons on the pump like "30E", "50E". You'd press one and then do exactly what you described with the difference that the pump stops at the preset mount. Other unattended pumps will ask you to put the card in, authorize a payment, the fill up and they will charge based on that authorization and all this using the regular ATM style keypad unit with a tiny LCD (the kind you'd see at any parking, or automatic ticket dispenser).
They are all very straight forward in their usage even when you don't understand a word from the written instructions. Very simple UI that's shared across industries to make sure people have some familiarity. With the downside that the UI looks very '90s, should anyone see that as a problem (Jalopnik probably would).
Not sure if it's across all of the US but I know in Hawaii I had to pay before I was allowed to put fuel in. Which made for this weird situation where I had to try and mathematically calculate how much fuel I needed.
And of course no tap/pay like the rest of the world.
At our (NZ) BP pumps we drive up, open the “BPMe” app, choose our pump number, enter a 4-digit PIN, fill up, replace the nozzle and drive off.
Another retailer (Z Energy) has a new “Fastlane” service where you drive up, fill up, then drive off. It uses a camera pointed at your vehicle’s licence plate to figure out who to charge. (I haven’t used this, though. I’m a staunch BP customer.)
My favourite UI mistake at a 24 hour self-service pump, a page on a 7" screen half way through the sequence of pages after selecting the gas type that simply asks for a Yes or No answer:
This small company did the software, probable not a lot of UX/UI experience there, so get something together that ""looks good"" and most importantly was done without extra cost (well, maybe GIMP or an "unofficial" PS version * wink * ) and voilà
The receipt screen looks nice because it's probably "native", meaning, there's some SDK for the ATM/dispenser hardware that makes it easy to ask questions on the interface
The thing that surprises me most not how bad each interface is, but that no 2 interfaces, even at the same chain or sometimes the same store, are ever the same.
On top of that, how can every one of those unique combinations be so consistently bad. Generally evolution finds at least one good option.
My local pumps don't have ads, but the number of questions before allowing me to pump are frustrating. "ZIP code?" "Carwash?" "Receipt?" "Rewards card?" "Debit or Credit?"
There's a gas station I used to live nearby, they have a free plastic glove dispenser glued to the side of the pump, so they you don't have to directly touch the pump
The graphics on the screen are honestly something I have never payed attention to. What really bothers me is the way the interface bounces you from touch screen to physical key pad a couple times.
At my local station, the final step before you actually get gas flowing is to push the "Enter" button on the physical keypad, which also happens to be right next to the "Call" button which if you accidentally hit it (wearing gloves, too fast, etc) you get the tinny disembodied voice asking "can I help you?" instead of gas.
I'm Canadian and I've never seen gas pump UX like this. It's mostly just terse instructions in a clear font on a black and white screen; certainly no cars represented.
Gas pumps, payment readers at restaurants and stores, ATMs, vending machines, etc...
I've always kind of thought it was because these important/widespread devices lacked sexiness. So you get engineers working on them and not UI folks. You get solid transactions, and UI is just not a priority.
Meanwhile at game companies, you get fancy UI folks self-selecting to work there.
It's crazy how much worse they have gotten. I used a decades old microwave. Physical buttons that made sense like (from memory): +30s, +1min, +5min, Low, Medium, High, Defrost, Cancel, Start.
A new one I used had a wheel to adjust value, and a touch screen to chose which value to adjust. The touchscreen was almost completely irresponsive touchscreen (like long hard press 5 attempts before it works). Effect was preselected despite rather than time which varies more. The visual indication of which one is being adjusted is ambiguous (one light, one dark).
Why isn't the PIN pad on a touch screen? Wouldn't that be more secure? Are people's hands just too filthy in the aggregate? I know those buttons feel pretty grimy.
They have been giving away paper towels and glass cleaner since the dawn of time. Why not hand out screen cleaners too, and let people wipe down the touchscreen if it needs it? People could also use them on their phones and would be delighted.
Or why not allow people to place their orders from their own screens?
There is no competition for this equipment around UX because there is little competition at all in this space. Its a remarkably regulated product, from weights and measures to EPA checks. UX only has to be good enough for the seller to close the order, and they go in 2500 locations. People buy gas on price and location, not UX, so the theory goes.
When working with credit cards and chip/PIN systems the entry of the PIN needs to be secure. This usually means the scanning lines from the keypad go directly to a security-hardened subprocessor inside the pump - the same one reading the PAN from the EMV chip or magstripe. Then the PIN/PAN block is encrypted and sent off to the application processor and/or bank to complete the transaction.
If PIN entry was offloaded to the application processor, that processor would need to be audited to make sure of certain requirements (PIN isn't sniffable, it isn't held in RAM after deallocation, encryption isn't breakable, etc). At the end of the day it pretty much has to be as hardened as the little subprocessor attached to the EMV reader, except any change to the app now needs to be recertified like the subprocessor is.
PCI is beginning to come around to allowing PIN entry on non-secure devices, see the following standard for more information:
PIN pads are generally designed so other people can't shoulder surf your PIN. Having it on the touch screen would make it trivially ready for everyone to see exactly what you are entering.
Also, unless someone physically replaces the PIN pad, which is difficult and likely to leave visible indications of tampering, then one can reasonably sure that entering your PIN is secure.
With a software touch screen you have no guarantee and no reasonable way of telling if the software is not harvesting your details and sending them off somewhere so your card can be cloned and used by criminals.
You mean when the article says "it has a user interface light years better than your gas pumps". That's not expressing time, it's expressing distance. Not sure if it's a common expression where you're from, but it's the same as saying something like "we're miles ahead of the competition." Distance ahead. So light years works as a substitute for miles.
Has anybody seen the "upgrade" that AM/PM (at least in California) to their pump terminals in the past 12 months or so. It is a breathtakingly poor UX and use of hardware.
I am pretty sure that the approximately 10 inch screens are color and reasonable resolution but the UI is entirely character-based and often manages to obscure what little text is presented by placing it close enough to the border that the outer frame easily hides it (at my height) leaving about 90% of the screen unused -- the easily visible part.
Not sure what happened on this project. I figured they would roll out v2 of the firmware/software shortly after I first saw it. Nothing so far.
I thought this was going to be about why diesel and gasoline don't have differently shaped nozzles, so you can't screw up your car engine with the wrong fuel.
Diesel instead of gasoline is almost impossible because the filler necks are different – at least where I live … Gasoline instead of diesel, however, is still possible.
My favorite screwup- my BP station prompts with something like "BP points 500, Yes No?" What am I supposed to answer to that? I just hit No on the theory that not doing something random is better than doing it.
The worst thing I've found about pumps with touchscreens is that half of them tend to be sitting in direct sunlight so they can't be seen. I can deal with crappy Camaro images, but if I can't see the screen it's 100% unusable.
