As designers, we appreciate great products. Designs that wow us with their experience, but also have that emotional connection of something beautiful and polished down to the smallest of details.
We smile at success stories of hard work, thorough research, beautiful UI, and perfect execution.
This is great design! We exclaim.
But we ignore one important part of the story. Most of those great designs are pretty modern, consumer-facing products that were created after UX became a thing. The visuals progressed the most in the last 15 years.
But the world…
The world is running on horrible, outdated, clunky user interfaces. Many of them coming from a time before anyone would consider User Experience a factor.
These apps run our railway scheduling, airplanes and airports, banking and financial system, sewage treatment plants, your insurance company, power plants, and more.
So while it may be fun to order a pizza with just three taps on a beautiful, ergonomic UI, the truly important things are very far from that ideal. They require a lot of time to learn, are often impossible to master and leave a lot of possibilities for errors — sometimes with grave consequences.
Before we get to the UI’s, let’s quickly look at the technology running a lot of these apps first.
Have you heard of Cobol? It’s a programming language created in 1959 that is used by most of the big businesses and governments to this day.
It is so rare and old-school in many ways, that there are jokes like:
The only time a company is looking for a Cobol developer, is when their previous developer died of old age.
And while apparently Cobol wasn’t all bad (according to some developers tombstone inscriptions), it’s a relic from the past just like those clunky UI’s of the ’80s and ’90s.
Yet they continue to run our lives in the most crucial aspects of them.
Y2K was short for Year 2000 and caused widespread panic because of the upcoming date change. The whole reason for that panic was coming from the fact, that developers used just six digits to represent a date instead of eight.
So 99 was simply 1999, but 01 was interpreted as 1901. The year 2000 would then be the same as 1900.
That could cause a lot of potential problems for most industries running these legacy systems.
As Doug White wrote:
Imagine [a] COBOL program that deals with county records to record births and deaths. If all the dates are stored as 6 digits soon you will have records which say something like 09/03/63. Now suppose, I live to be a hundred years old, my birth is recorded as 09/03/63 and if I die on my birthday 100 years later my death would also be 09/03/63.
Many companies frantically worked on a fix in the late 90’s, while the media painted scenarios of airplanes falling from the sky and powerplants exploding.
Luckily almost none of that happened.
UI should not be easy
Before the UX revolution, the idea of an easy-to-use interface in a “Pro” app was laughed at. After all, if someone’s a “Pro”, they surely can learn how to operate the heavy machinery of a hard to grasp, 500 field form with many conflicting labels.
An old Windows XP meme, but in reality many error messages were just as helpful.
An old mentor would sit down with aspiring “Software operators” and while stroking his long, white beard he’d guide them through the interface.
This field says “the amount” but it’s NOT the amount. It’s only used for a scenario when the transfer is going in the internal test network. For real amounts just use the field “AMNT2” — that’s pretty self-explanatory right?
Clunky UI like that led to Citibank sending people $900M instead of $7.8M in interest payments. And that transaction was confirmed and done by three people.
There was no help and no guide, apart from things people remembered from onboarding meetings — often done twenty years prior.
If it ain’t broke don’t fix it
Some large corporations still stick to this rule, as rebuilding their legacy system could cost Billions and can lead to a set of completely new problems.
But losing money like in the example above is a good motivator for change and there will be a rise in very large ground-up redesigns of all that important software coming in the next few years.
Let’s hope we, as designers, are ready for it, as the level of complexity is a bit higher than a scooter rental app, and the consequences of errors are a lot more significant.
I personally worked on two products like that (in fintech and banking), and I believe they require a completely different mindset and a profound level of understanding before any “design happens”.
It probably is not going to be as simple as that.
More UX work
So while we may feel that the UX industry has reached its peak in 2009–2012, it may actually have a revival soon.
While those simple products we’re mostly working on right now can often be a mix of heuristics and a nice UI on top, these legacy apps are on a whole other level of complexity.
If you’re a problem solver, you should be excited! There will be a plethora of problems to solve soon, with very high stakes and an immense level of responsibility.