Taking me between the wrong two cities was only the first of many disappointments I found about the Auto Train. On booking my ticket on the Auto Train they stressed how important it was to arrive at Sanford early. How early, I asked. Four hours early is what they suggested. "And you can eat your lunch at our nice facilities," they said. Fortunately, friends had warned me that the food at the train station is scarce, mediocre, and overpriced. They turned out to be right, but I bought some ice cream there anyway to help alleviate the Florida heat and pass the time, since we had left West Palm Beach at 9 a.m. to get to Sanford at noon to catch the 4 p.m. train.

Arriving at noon, there was already a line of cars waiting to board the train. It turns out that it is important to arrive early because loading the cars on the train is done at a glacial pace. Nonetheless, at 3:30, minutes before loading stops, cars arrive and are still accommodated at the last minute on the train. Mine was among the first cars to get loaded. While my early arrival made things convenient for the Auto Train, I had no idea how badly I’d be punished later for helping them out. Arriving at Lorton, I learned that the unloading process is almost as slow as the loading process and that the first cars on are the last ones off. Two hours after they started unloading cars I was finally reunited with mine. And the unloading process is designed in such an interesting way— you are required to stand outside and watch every car (over 200 of them in my case) unload until you get yours. Auto Train’s implicit strategy is to reward their worst users and punish their best ones. Folks who squeaked in with their cars at the last minute saved hours of waiting at Sanford and then saved hours more at Lorton as their cars came off first.

Another seemingly nonsensical process: On loading, your car is assigned a random number that is magnetically attached to the door of your car. If the car before you gets number 304, you might get 259, and the car that follows you might get 007. You are told to be sure to remember that number. When cars are unloaded, they announce the number you had to remember. If your car is 259 and you hear car 258 announced it tells you nothing about when your car will be unloaded. So it is necessary to stand nearby and simply wait until your number is called—a process that can take hours. Cars are placed in a small area for you to pick up. If people fail to take their cars out of the small area quickly enough, the entire unloading process slows to a crawl and would soon stop. Is this the best way? Would it be hard to announce cars by the names of the owners, the kind of car, or their license plate numbers? Would it be hard to assign sequential numbers? No, there are many things that could be done to make this process easier for users. But as it is, the process is quite easy for the Auto Train. And to accommodate the long unloading and loading times inherent in this process they just schedule the train’s departure many hours after its scheduled arrival.

Getting back to the Sanford-Lorton issue, we might question why the Auto Train doesn’t offer service between the cities people actually want to travel between. It turns out the reason is that the three-deck car carriers they use cannot fit under the bridges beyond the cities it serves. Then why don’t they use two-deck car carriers? It would make the train longer, and maybe heavier, but long, heavy trains should not be a problem. In fact on the same tracks that the Auto Train uses, freight trains much longer and heavier than it travel freely. However, the Auto Train is as long as it can be. Its length is limited by the length of the sidings that it must pull into to let freight trains by (since freight trains are too long to fit into a siding). Furthermore, freights all have priority over the Auto Train. Oh, and did I mention that there is a single track which must accommodate two-way traffic from Sanford to Lorton? And sidings are spaced fairly far apart. The result of all this is that the Auto Train does not go to the places people really want to go. Though it never stops at any station on the way, it does make many long stops at several sidings adding enough time to the trip so that any car on Route 95 could easily beat it—even counting the time spent at rest areas. Given that the Auto Train makes many stops at sidings, why do they advertise the train as non-stop? Well in railroad jargon, a stop is a stop at a station to let passengers on and off—and it never does that. To a user of the system, of course, a stop for a half-hour to sit on a siding is as disruptive as one at a station, but that doesn’t seem to count.

The Auto Train has many system constraints. It has lots of tradeoffs to make. If it makes its trains longer it may have to increase the length of its sidings. To travel to New York it has to raise lots of bridges or lower lots of car carriers. Anything it does to offer its users the service they need will be quite expensive. It has chosen to do none of these. Users, it has decided, will fit into the system as it is.

This is not to say that the Auto Train is not addressing some needs. It is spending substantial sums of money to upgrade its facilities at Lorton, Virginia to make waiting more pleasant. But they are not solving any real problem any user has. In fact they are overlooking user problems and just making a bad situation a bit more tolerable. No user wants to wait in Lorton to board or unload a train. Users don’t even want to be in Lorton. Users don’t want the wait to be more tolerable, they want it to be shorter and they don’t want it to be in Lorton at all. But those are hard problems. It is easier to solve easy problems, even when that is not what the user really needs.

But what about the reclining seats that you can lull yourself to sleep on? This turns out to be just another empty promise. The seats don’t recline enough and pairs of seats have hard ridges between them that makes sleeping across two seats nearly impossible. The Auto Train promises nirvana, but delivers far less.

Bored with the promised wonderful scenery that consists of the shacks of poor folks and the declining industrial plant that one finds along railroad tracks, I tried to go off and watch the movie in the movie car. But it is on two tiny TV screens at either end of the bar car where the lights, laughter, and general chaos make seeing or hearing the movie all but impossible. It didn’t matter much though, because no seats were available anyway.

Overall, I concluded that the only reason the Auto Train is so successful—it is hard to book a seat during busy times—is that it has no competition. Having no competition, it seems that the folks who run the Auto Train have endeavored to offer a service completely insensitive to users’ needs. But this is shortsighted at best because in fact, there is competition for the Auto Train. At least twice a year Route 95 is filled with Lincolns and Cadillacs full of gray-haired seniors making their semi-annual migrations between the northeast and Florida. Hordes of trucks transporting the cars of other people unwilling or unable to drive also vie for space on Route 95. And others arrive by the planeload or are transported by friends and relatives. The Auto Train for many is simply the least of many difficult ways of getting where they need to go. Because so many people make the trip, the Auto Train is usually able to fill one of its trains each day with the small percentage of folks willing to tolerate its shortcomings.

If the Auto Train offered really good service at an affordable price it would be swamped with customers. That would mean it would have to run more than one train per day each way. Isn’t it inconvenient to have just one train each way per day? Sure it is. But for the Auto Train folks it works out well since it is the same train going back and forth. Think of the cost and complexity of running more than one train. It boggles the mind. And at any rate, could they really find two trainloads of people who wanted to travel between Sanford and Lorton? Expanding a bad service is unlikely to bring in more customers or to make them happier.

But there would be great benefits to having more people use the Auto Train. Carrying cars and people on one long train is environmentally much more sound than having those cars and people drive individually. Getting lots of cars off Route 95 and onto trains would clear much of the road congestion. If it were easier to get between Florida and the northeast, more people would do it, adding economic benefits to the region and elating travelers who would otherwise be unable to make the trip. But the Auto Train is unlikely to change. Why should it? It has all the business it needs and it owns the only rails that could be used for the service.

After finally retrieving my car, I drove north on Route 95 towards Philadelphia dreaming of the kind of train that I would really liked to have been on. It made me also think about all those users—computer users, that is—who dream of the kind of IT organization they’d like to take them where they really want to go. If your users are like this, maybe you have them on an IT Auto Train. The Auto Train is actually a great idea, but it’s being badly executed by people who have lost sight of a few principles:

Act like you have competition—you just might. Our IT organizations seem to have exclusive ownership of the rails necessary to provide IT services at our institutions. And we no doubt have all the customers we need—in fact, maybe more than we need. A campaign for good dental health used the slogan "Ignore your teeth and they will go away." That may eventually be true of Auto Train users as well as our IT users.

Don’t punish your best users and reward the worst ones. Do you ever give the "squeaky wheels" great service, but leave the nice folks who don’t fuss too much to fend for themselves? Do you spend more time with folks who have never bothered to learn your systems and less time with those who have?

Don’t train your users to learn complex systems. Fix the systems, not your users. Give users the simplest procedures possible; not the simplest they can handle. IT organizations need to look at every process and procedure from a user’s point of view. It isn’t enough to decide that users can handle something. Users have lots of other things to handle and doing IT shouldn’t just be within users’ cranial capacity; it should be as easy as we can make it.

Don’t provide a service that takes your users only a part of the way. Don’t leave them floundering to go the rest of the way on their own. Change your infrastructure to accommodate your users.

Be user-centric. Do what’s best for your users, not your systems. Make your priorities the same as your users. Consider how a user will use a system, not the "right" way to use it or the way you would use it.

Don’t expand a bad service. Fix it first. Don’t make it easier for your users to tolerate a bad situation. (And don’t fix it by just documenting it.)

Talk to users in terms they understand. Talk means written talk and on-line talk too.

Don’t lie to your users. Not even white lies. Truth (in fact and in spirit) in advertising is essential. Don’t promise what you can’t deliver.

Help your users get off the IT Auto Train. Start offering stellar service. The benefits will be enormous and university-wide. Opportunities lost by not employing appropriate information technology will be saved. New efficiencies will be discovered. Disgruntled users who have given up on projects vital to the university will be reenergized and those projects will move forward. Create an IT department that takes people where they really want to go.