Nothing affects usability more than the speed of the system and interaction between user and machine. In a technology led industry often we sacrifice speed all to soon by selecting the latest emerging "cool" technology and throw away the best chance of producing a truly usable system.

During the past 5 months, I have been working in Ireland and living in Scotland on the weekend. My weekly commute between my home on the West coast of Scotland and my apartment in Ballsbridge, Dublin takes just over 2 hours at best and 3 and a half hours on a bad day. I take a half hour flight from unfashionable Prestwick International airport to Dublin with the basic no frills Ryan Air.

There are several alternatives for me when traveling. There are numerous ferries from Scotland to Ireland. I could for example journey from Troon to Belfast and on by car or train to Dublin. The total time would be around 5 hours on a good day. Rather slow and the cost perhaps slightly more than flying.

The most obvious alternative would be to fly from the much larger Glasgow Airport to Dublin using the Irish national carrier Aer Lingus. A traditional airline offering a high standard of service and modern aircraft. Glasgow Airport is equi-distant from my house with Prestwick. So why do I choose to fly from sad, somewhat delapidated (but improving) Prestwick using a low end airline such as Ryan Air which offers a coach class service in very very old aircraft?

The answer is simple! Speed.

The Journey

It is easily the fastest route. My goal is to get to my destination. My task is to travel. My aim is to achieve the goal in the fastest time with the minimum of stress. Only Prestwick to Dublin offers this. How?

Prestwick airport is situated at the end of the A78 trunk road. To get there I turn left out of my drive way and proceed down the A78 for around 20 miles. Most of the route is Motorway class road. I can get there in around 20 minutes.

Once there the car park is close to the main road, and close to the terminal building. Prestwick is a small airport which handles a small quantity of traffic. Parking is easy. Walking to the terminal is quick and easily navigated.

Checking-in is a pleasure. There is often no queue. Check-in can be 20 minutes before take-off.

The flight is quick, only 30 minutes and usually runs on time, arriving early.

At the other end, Dublin is busier and more congested but it is still a short walk to the rental car desk, and a short walk from there to the pickup point.

The final leg, driving through Dublin is a chore but on a quiet Sunday evening its no problem. I am home only 2 hours after leaving home at the other end.

Alternatively, I could go to Glasgow on a busy, narrow winding road, often packed with commuter traffic or slow moving heavy vehicles and farm machinery. The parking lots are vast and crowded, finding a space can be difficult, the walk to the terminal is longer, the check-in queues can be long a peak times. In comparison though, Glasgow is a feature rich airport with good shopping, a nice range of cafes and restaurants and other services.

The Lesson

The Lesson to be learned here is that the over-riding factor in product or service provision was Speed of delivery. The journey is stressful. Reducing the time and easing the friction along the way is important.

My goal is to get to the destination. I do this by making a journey. There are many factors involved: Road to the airport; ease of parking; journey time from car park to terminal; check-in process; airline; aircraft; flight time; ease of rental car check-in; journey time to the rental car pick-up; road from the airport at the other end.

This provides an interesting example, as no one provider for air travel service can control my decision. Aircraft designers can design a nice airplane but it won't make me use it. An architect could design a nice airport but it won't make me use it. Fast roads provided by the local authority won't make me use them either. A nice service and an appropriate flight time from the airline, won't make me use them. It is the whole chain of services related to achieving my goal which affects my decision. The "whole product" which delivers my goal is a road network, an airport, an airline, an aircraft, a rental car company and another road network.

Making Software Solutions faster

The software industry has an advantage over the air travel industry. Providing a "whole product" solution is (often) within the grasp of a single provider. Using software is stressful. There is a lot of what Alan Cooper has termed, "cognitive friction" involved. Delivering a fast system which minimises that friction is the best route to true Usability.

The big lesson from my weekly commute is that there are many aspects involved in providing a solution which allows a user to achieve a goal. That system is complex. It involves an underlying architecture, network infrastructure, database, business logic, interaction design, and often physical environment design.

At all stages in the architecture and design of a system, SPEED, must take a parimount consideration in choice of solution.

There are IMHO two keys to this. Fast technology and Good Interaction Design.

Fast Technology is often older and warmer

The choice of technology is often made based on what's new and what's cool. For example, Swing/JFC. Well frankly that might not be the best choice because it is sooooo slooooowwwwww. It might be better to choose something a little older, a lot warmer and perhaps rather grey and boring. Why? Because it probably runs a lot, lot faster.

Too many times, we throw away our best chance of usability by picking a new technology for some other supposedly "compelling" reason. When C++ was new, C was still a lot faster at execution time. When Java was new, C++ was still a lot faster at execution time. When JFC was new, AWT was still a lot faster at execution time.

So what am I saying here? Am I advocating a return to assembly language programming? :-) No! There are still many factors involved in a technology choice, but I feel it is the duty of the Interaction Designer to strongly advocate, older more established technologies when there is a real speed advantage. Usability has to get a voice at the table when the choices are made.

It is often argued that 2 years from now (or 5 years) everything will be OK as hardware will be faster. BUT, what about the poor users who have to put up with diabolical usability until then?!

It is really important to strike the correct balance with the maturity of the infrastructure technology. If it is in its infancy, the chances are that the vendor hasn't been making an effort to deliver performance. Two choices emerge for the application developer. Apply pressure to the vendor of the infrastructure to deliver strong performance soon, or choose an older, perhaps less hip, less functional alternative that delivers performance now.

Delivering IT Solutions in a NOW thing, not a tomorrow thing!

Fast Interaction is minimal Interaction

Minimising the hassle for the user is another important aspect of speed. Reducing time delays on response is important but equally important is reducing the number of dialogs, key presses, mouse movements and cognitive friction for the user.

As an Interaction Designer it is important to challenge every attempt to add a key press, a mouse movement, a selection, a button click, and most of all, a Dialog or Wizard into the system interaction design.

It is important to communicate this with your management and get buy-in to spending more time on the interaction reducing it down. Metaphorically, you have to put the car park close to the terminal, make check-in easy and fast, put the gate close to the check-in desk, minimise the distance to walk, make the car rental desk obvious, and put the collection point nearby. Don't put unnecessary hurdles in the way. Don't ask for unnecessary forms to be filled in, don't ask unnecessary questions. Be polite. Learn about your customer (or user) and predict there goals.

Every time you add on an unnecessary interaction or fail to provide a sensible default action, you are puting another small hurdle up for the user, adding other few seconds to that journey towards the goal. Add enough of them and you might just tip them over the edge and your user goes elsewhere. You just lost the customer!

How much did that cost?