Monday, July 9, 2012

Airline UX is terrible, but website design is not UX

We all describe, or gripe about, what we know. Many UX folks on contract or who speak often, like me, spend a lot of time on the road. We get used to the stupidity of airports, airlines, hotels and generally travel. 

But really, airline booking is one of the stupidest, most jargon-laden, internal-process-driven things I deal with on a remotely regular basis. I have to go to sectors like health care (maybe worse as they are supposed to be "patient focused"), before I get to similar levels of difficulty.

Periodically, I see some clever designer or other posit how the process of airline checkin could be improved. But I haven't seen a one that would make a lick of difference. I think we're miles from what happens when you click a tab, or remembering my previous settings better. And in doing so, there may be a lesson in the difference between interface and interaction design, and the concept of designing for user experiences. Let's see. 

First, let's baseline the issue. If you don't travel much, or don't think about it, the basics of air travel are:
  • You have to pick dates you want to travel.  
  • You pick an airport. By IATA code, instead of city or airport name, makes it easiest on the system. But don't use ICAO codes. 
  • You pick some other constraints, maybe, like how many connections you can put up with. 
  • Sometimes you can pick times of day. The results are not clear, and it's usually departure time, which may not be relevant to you. 
  • You get a list of flights. They are delivered with codes and dates and lots of other data, whose main point of reference is price. Then you parse by time of departure, time of arrival and total flight time yourself. If you can figure it out.
  • Let's not even talk about getting cars, arranging shuttles, or hotels. 

But what about that do I as a traveler want? I certainly do not care about airports. Really. And it's much worse, as when I fly to San Francisco, as I am doing today, there are three perfectly reasonable airports. A few travel sites allow you to "also search nearby airports" but only a few. And you still have to start with an airport code, or I guess a major city. 

Or, take this other good problem I encountered a few months ago. We're at an event in what you might call "the middle of nowhere" in Oklahoma. People drove from all over the country, some flew, and more than a handful came from overseas. They all looked at the map, followed the typical method, and picked the nearest (by their perception) big airport. Mostly Kansas City, as we could pick them up and shuttle them over, but a few from Tulsa. 

However, while driving down, we see billboards (a bit too late) and I check. Not only does Joplin, MO (maybe 25 minutes away, vs 2-3 hours to Kansas City) have jet service, but it's significantly cheaper to terminate your travel there. 

So, we see your requirements as an end user are:
  • You are (or know on a certain date you will be) somewhere. 
  • You want to be somewhere else, by a specific time.
  • As efficiently, and cheaply as possible.

I may be more annoyed by this disconnect, because there has been a quite well thought out dream of totally changing air travel. 

William Langewiesche (I think, I cannot find it listed) in the late 1990s wrote a great article putting together some of these thoughts, centered around Free Flight, and some related concepts with the coming of cheap, small jets. Think the Honda Jet. This, clearly, never came to pass and I am not sure if it's just inertia, or 9/11 security stupids got in the way, but even 15-20 years ago, this was possible. Today, it's even easier. 

Imagine you want to book travel tomorrow. You simply go to the internet. 
  • Say where you are. Or will be, on the date you want to travel. 
  • Say where you are going. 
  • Say when you want to be there. Add other constraints, if you have limited time on the other end as well.
  • That's it. 

The response you are offered is a single choice. But, it meets all your requirements, within maybe an hour or two of when you requested. Show up at a small local airport, probably much closer than the major international hub somewhere in your city, at a specific time. When you get there, you give your boarding pass over, and walk to a tiny plane. I mean, maybe 6-8 people. Maybe 25. But probably not bigger. As it's a ride-share system. You and all the people you are flying with want to be all are from very near this airport, and want to go within a few miles of the destination airport. 

No layovers, no stops. And this would be happening many thousands of times a day all over the country (international may be trickier with just over-ocean operations). We're well into the era of business intelligence, but even giant hub-and-spoke systems are getting in the way of applying these usefully. Things like Free Flight, could make point-to-point travel like this even more efficient. When your plane is circling to land, or you sit on the plane waiting for a gate, this is not even analogous to a packet collision; it's more like a fast busy on your home phone. 1930s queuing and traffic management theory are still being applied and getting in the way of efficient transport. 

Right now, all airplanes travel on "highways in the sky" which pass through all the major cities. Just like you can't drive straight cross-country but have to take the interstate wherever it goes, airplanes make lots of turns to get where they are going. We don't need to. We need a shift as radical as that to packet-switched networking. 

Back to the booking experience, we can tie in things like intent. On this trip, I am going to an office building, and don't own an apartment or anything so will be quite willing to put that address in, and click the checkbox, "find me hotels" which becomes a more fluid part of my "which airport to fly into" parameter. 

Even without that ideal state, relatively simpler changes to the way booking systems work would seem easier. Just take the concept of starting and destination addresses instead. Actually, as I type this, I can think of a new business for Google to get into: 

Select the airplane, and you get driving directions to the airport, then flights that meet your needs. Presumably, this would only appear when the trip makes it reasonable. There's no airplane directions to the nearest coffee shop. 

Of course, Airlines could improve it even more by implementing other services, like shuttles to and from the airport. This always bugged me. What if instead you book a flight (even with the current stupid system), and they say "be ready at the address you gave us at 10:15, we'll pick you up." A Delta van with a guy wearing a Delta Airlines van shows up at your house. The van driver would dump your checked luggage right into a conveyor when you get to the airport, and you go straight through security and to your gates instead of the cattle call of "show up at the airport at least 2 hours early." 

Besides desperately wanting this for selfish reasons (I book a lot of travel), it is a good case study in the scale that UX design should consider. UI changes, even radical ones, can only have so little impact on the experience if the business process doesn't change. And small business process changes may not be enough to really satisfy your customers or differentiate your business from the competition.

While improving air travel is not something that I can see any startup working on tomorrow, think about these same demands and constraints on your business. Will "improving the UX" or"mobilizing the business" really matter to the end user? Could stepping back further, and considering larger or tangential solutions for your domain or the whole enterprise help solve issues instead? 


Will England said...

Best comment of the article?

"UI changes, even radical ones, can only have so little impact on the experience if the business process doesn't change"

Ah yup. 11 years on the front end of the biz doesn't make a dang bit of difference because the back end is still broken. At least we don't swivel-chair orders from system A to system B anymore.

Change the core, the UI follows.

Make the core do what your *customers* demand, you'll be successful.

Steven Hoober said...

Ah, manual processes. Why do you think I always include lines about "stuff written on Post-Its" in my design process presentations? Because I have been to your call centers :)

Do you remember the ad, like 10 years ago, with an order processing (or something) box. Put paper in one side, comes out the other. Inside the box: people. Perfect representation of the business-understanding of so many processes.

smith said...

Amazing how simple it can be to communicate with people and have them understand a certain topic, you made my day.

oklahoma web designCity

Webmarketing Experts said...

Thanks for sharing this great content, I really enjoyed the insign you bring to the topic, awesome stuff! review

Unknown said...

I am the real estate field. I have read your tips which is very good to the real buyer as well as seller.

Web design San Francisco

Tbcreation said...

You made some good points there. I did a search on the topic and found most people will agree with your blog.
Graphic Design | Graphic Design Rochester NY

Publish your passions said...

Positive site, where did u come up with the information on this posting? I'm pleased I discovered it though, ill be checking back soon to find out what additional posts you include. website improve