Approaching the gate for my flight from Dublin to San Francisco, I thought I must have misread the boarding gate listings. Though the flight was scheduled to depart in under 60 minutes, not a soul stood in the lounge area except two Aer Lingus gate agents.

I knew I hadn’t missed the flight. Where were all the passengers?

Rhetorical question. I knew exactly where they all were: hostages to well-intentioned technology back in the immigration area, that's where.

Yet again, a major IT project intended to alleviate problems had instead exacerbated them (IT failures, thy names are legion). In this case, it was the US customs check-in system introduced a while back, which steers US passport-holders to self-serve touchscreen kiosks that scan and process their passports, clearly with efficiency and better service for travellers in mind.

Anyone who has headed to the United States from Dublin Airport over the years will be aware of the problem the kiosks hoped to tackle. At certain times of day – as US flight departures cluster across the morning, especially during the tourist season – the queues in the US customs and immigration pre-clearance area would form a slowly shuffling, tightly undulating river of frustration.

The logic of the kiosks, I suppose, is that if people got the passport processing and photo-taking bit done before entering that final snaking queue to the customs agent, the agents would only need to spend a moments rather than several minutes with each individual.

While this shift in processing works – or at least, adds no delay – when passenger numbers are moderate, it fails when numbers build. Moving the time-consuming element to kiosks at the entry to the customs hall creates new queues out into the security area (US passengers all pass through a second security zone before customs).

Security concerns

Someone decided that this issue could be tackled by putting more kiosks upstairs, before the descent to customs. But on one busy day this week, floor staff had to control the numbers accessing the upstairs kiosks, and then, restrain the overflow numbers. The added queue of those who had been to the upstairs kiosks and were now waiting downstairs, beyond security, in a separate line to go through customs, created such a human logjam that staff had to stop people from descending the stairs to security – creating another queue upstairs.

At one point, security screening came to a halt, the conveyer belts turned off, because the queue to get into the customs area wrapped around the security area floor. And to make matters worse, the computer system had gone down in immigration for a period shortly before, halting everything.

Staff had to start directing US passport holders through the shorter customs queue on the non-US passport side (without using the kiosks). So, ironically, the fastest way to get through customs became the old fashioned one of simply being directed to a customs agent and avoiding the new technology kiosks altogether.

The end result of the supposedly efficient kiosks upstairs and downstairs: frayed tempers and, inevitably, delayed flights (like mine).

And that’s all why, at the point where my flight should have been well into the boarding process, the gate was deserted and more than half the passengers had yet to arrive from customs.

Did anyone designing this new system consult those frontline workers – the customs agents, or the security or airline staff – on what might work best before it was implemented?

All of this mayhem, which must be a daily nightmare for staff, is a textbook example of how technology offers little benefit when implemented without getting adequate user input or considering in full the context in which it will be used across a broad range of user scenarios.

In theory, the kiosks are a fine idea (on a day of moderate traffic and if US passport holders don’t make up the bulk of passengers). But they just do not work well when passenger loads are heavy (read: the whole tourist season), flights are closely spaced, and given the current configuration of the Terminal Two building for US flights, which creates several bottlenecks.

Even if, as is possible, this new system means passengers spend about the same amount of time queueing as under the old system, the perception is of sluggish chaos, thanks to those bottlenecks and the uncertainty of passengers – especially tourists – as to what is happening, and why they are being herded through so many queues.

The level of exasperation and confusion was as high as a transatlantic jet. Not good for tourist perceptions. Sacrificing end users to a nice concept is never good for any IT project.