Coda to What’s Our Job Here?
In my last post, “What’s our job here?”, I proposed a different way to assess interaction design problems. In that post I hinted at, but didn’t actually describe, the actual, quotidian solution to the interaction. I will do so here, just so you can see how I think about this stuff.
Here’s the design problem my colleagues posed to me:
Shannon and Steve were in southern California teaching a Cooper U class. They discovered a new automated toll road in Los Angeles that uses wireless transponders in each car instead of having toll booths. The problem is what to do when someone gets on the toll road without a transponder. Currently, the driver is liable for a costly penalty unless he goes to a hard-to-use website and pays.
Shannon and Steve were in a borrowed car when they accidentally ventured onto the toll road, and not in a rented one, which would be the more typical case for people visiting from another area.
The conventional approach.
When you look at this problem conventionally, that is, from the point of view of the toll road needing to get every penny it is due, you fail miserably.
The simple, obvious, and wrong solution is to send a dunning notice to the owner of the borrowed car. Cameras in the toll booth read the car’s license plate, and thus identify the owner, if not the actual guilty party driving it. A dunning letter to the car’s owner is certain to generate ill will, likely to provoke bad behavior, and in most cases it will ultimately fail to collect the money owed. The owner rightfully doesn’t feel that the debt is theirs, having merely done a good deed in loaning the vehicle, and will resent being penalized for their generosity. Whether they pay or not, such a demand will be greeting with irritation and resentment.
What’s more, the owner will quite likely just ignore the bill and there’s only a very small chance the city will ever actually collect. The only recourse the city has is to initiate collection procedures, and there is no conceivable way that the six dollar toll is worth the cost and hassle of collecting it.
If there were some way for the city to identify the actual driver, and dunned them, even after they have returned to their home state, there would at least be some defensible rationale, as the city is now billing the actual violator. But again, this is not really a collectible bill. Who would bother to mail off six dollars to a city visited only rarely for driving on a road unlikely to be driven on again? The chance of collection is dim.
The goal-directed approach.
Alternatively, by examining the problem from the point of view of the driver instead of from the toll road, the solution becomes clearer.
Shannon and Steve were visitors from far away, in a borrowed car, traveling on that road for the first time, and likely for the only time. What they want is for the problem to just go away. It’s way too much overhead to set up an account with the city to pay a single toll. They might not have even used the toll road if they realized there was no way to pay cash. In this first inadvertent learning experience, Shannon and Steve want the toll to be waived, and that is exactly the correct solution. The city should cheerfully waive the toll.
The system can recognize that the car had never been on that toll road before, so it can easily let it go without exacting a toll or levying a penalty. It can also easily take note of the license number for any future transgressions.
If the system assumes this was a simple mistake made by out-of-towners, it fails safely for everyone. Shannon and Steve are not penalized and they are not made to feel guilty. The car’s owner might be alerted with a letter or email, but no action means no resentment. The city loses a minuscule bit of revenue, but they gain an enormous amount of goodwill and future awareness. Everyone feels that the system was fair and reasonable. Everyone resolves to do better next time.
The system only needs to act forcefully in clear cases of cheating. These are easy to identify because the same vehicle incurs multiple violations without a transponder. For the city, this willful bad behavior is worth pursuing, both for the money and for maintaining order. For the driver, this is justice.
The fact that Shannon and Steve were in a borrowed car and not a rental car means that their situation was atypical, as most out-of-town drivers would be in a rented car. Car rental agencies already have a relationship with the city, so in this far more frequent middle case, the fleet is already transponder-equipped, and the toll is easily tacked on to the renter’s bill. Simple cooperation between the toll road’s and the rental agency’s computers can make this a trivial transaction.
Cooper’s design methodology is called “goal-directed” because we know that design problems are much easier to solve when you look at things through the user’s eyes, understanding what their goals are.