The government’s legacy IT systems have been in the news recently. There’s a great set of articles by Jordan Press, from November and February, and a follow-up radio interview that’s worth a listen. Within the government, as these articles show, there’s a growing concern that these systems – software code and mainframe computers that underpin critical services and benefit programs for millions of Canadians – could fail unexpectedly at any moment.
This isn’t a new problem:
Many of the information technology systems that the federal government relies on to deliver programs and services to Canadians are aging, and several are at risk of breaking down.
Even if systems are currently working, a breakdown could have severe consequences. At worst, some government programs and services could no longer be delivered to Canadians.
That’s from a 2010 news release from the Office of the Auditor General. Practically all of the systems identified as at-risk a decade ago are still in use today. Amid other government priorities, IT systems have never been high on the list. As the United States’ 2013 Healthcare.gov crisis shows, though, we live in a world now where technology successes and failures are directly connected to political successes and failures. Outdated and poor-quality IT systems can constrain political leaders’ ability to achieve their goals, by delaying or preventing the successful rollout of services that their citizens are counting on.
A “lack of focus on end-users”
The complicating factor in discussions around legacy IT systems (and their need for replacement) is that many of the services that these systems support don’t work well as-is. From the perspective of a Canadian trying to apply for a program or benefit, they’re often confusing, hard to navigate, easy to complete incorrectly, and involve several stages of printing off and mailing forms or dropping them off in-person at a front-line office.
…Few of the online services offered to Canadians by departments are client focused. For example, we found online services that were difficult for users to navigate or that could not be completed from start to finish through the online channel only.
…the perception of the service is very different depending on whether you are talking to the service provider or to the citizen trying to navigate the red tape.
Over the years, our audit work has revealed government’s lack of focus on end-users, Canadians.
Government… needs to be good at service delivery to remain relevant.
In short: our difficult to navigate, user-hostile services are the ones at risk of failing. Our user-friendly services by and large don’t exist.
Two questions that are actually the same question
The maintenance and repair (or replacement) of IT systems is often treated independently from the services (for example, a Canadian applying for a benefit program) that they support. These systems might directly power online forms and interfaces, or, depending on the case, they might be several steps removed – storing data and calculating eligibility criteria for requests that come in through phone or paper channels.
The divide between conversations about systems and conversations about services is a good example of Conway’s Law – that organizations create products and processes that reflect their internal communication structures. In this case, siloed IT units are responsible for system maintenance and upgrades, while the design of programs and services happens in other parts of a department.
For a person using a service, though, the distinction is meaningless. If a service is beautifully and intuitively designed, but the systems it runs on are down for maintenance, it’s useless. If a system is up and running, but the service is a long, burdensome form that people stop filling out halfway through, then it isn’t helpful either.
The two questions worth asking are: does it work?, and, can people use it successfully?
As the risk of failing IT systems has finally started to be recognized across government, I’m worried that people will only ask the first question, and not the second, more important one.
(“Successfully”, it’s worth saying, is doing a lot of work in that sentence. How many people can find and use a service? Does it take them an unreasonable amount of time to get through? How likely are people to make it all the way through to the end? Are particular groups of people – using accessibility tools, for example, or on low-bandwidth network connections – more likely to find a service difficult or impossible to use? Each person’s determination of whether a service is successfully usable or not can vary.)
IT administrators are often content to stop at, “does it work”. Replacing legacy IT systems is difficult, and it might seem easier to make the bare minimum of changes (porting a mainframe system to a slightly less old mainframe; emulating it in an expensive, proprietary emulator for 1960s hardware, etc.). Doing that may keep a system working for a bit longer, but it doesn’t make it possible to make the services it runs more user-friendly, more flexible, or more adaptable to policy and program changes that political leaders may want to bring about. Investing in more foundational changes that make it easier to change, more quickly, in the future can be a hard sell, but worth it.
The test for public services in 2020
Surprise: it’s actually the test for public services in 2014 but we’re just catching up. The test is, can I complete this service online, on my phone, from start to finish, at two in the morning? The inspiration is the UK Government Digital Service’s 2014 redesign of the Carer’s Allowance, a UK government grant program for people who care for elderly or sick family members. After the service was made available fully online, GDS found that there was a daily spike in traffic around 1 or 2 in the morning, when people who had been caring for their relatives all day finally had a chance to apply for the program. If they had to phone a call centre or apply at an in-person office – as was previously the case – they would have had to take time away from caring for their loved one.
In 2018, a panel of Canadian technology industry luminaries wrote a report about making Canada successful in a global digital economy. It includes a challenge to the Government of Canada to digitize every public-facing service and make it available online and mobile-friendly by 2025. Echoing the Auditor General’s reports, they write:
Every government service needs to be delivered online to eliminate the need for in-person visits to government offices and the printing, scanning, and mailing of government documents.
The way our government is currently delivering services to Canadians is no longer appropriate for our times.
Doing this in practice requires changes both to how we think about managing IT systems, and to how we design services themselves. Ultimately, those are two parts of the same question, and whether we succeed or not has profound consequences for people’s everyday lives.
Why this matters
Making services that people can actually use matters, because it’s the most vulnerable people in our society who get burdened with the most bureaucracy. This work matters, because people aren’t receiving the benefits and services they’re already eligible for. It matters, because – although the government does a vast range of important things that aren’t service delivery – services are what people see firsthand, and the quality of the services they receive can directly affect their trust in government. It matters, because a whole range of middlemen – from tax advisors and immigration consultants and customs brokers, to Member of Parliament offices running passport clinics, to software companies advocating to be the front door to government services – are a symptom of the ongoing failure of the government to provide services that everyday people can successfully use on their own.
We owe it to everyone to build public services that they can use. I’d like to amend my test criteria to add: a good public service is one that I can complete online, on my phone, from start to finish, at two in the morning, in less than 10 minutes. At my old company, one of the hardest things was convincing clients that the 30 minute research surveys they wanted to do were too long. No one would make it to the end. If you want people to successfully complete something (a survey, a form, an online transactional service) you have to be conscious of the effort and cognitive load you’re putting on them. It’s a big part of inclusive design, which is finally starting to get some important recognition.
How many government services meet that test: fully online from start to finish, that you can do on a smartphone any time of day in less than 10 minutes? I can only think of one, out of the more than a thousand unique services the federal government provides.
We’ve got nowhere to go but up.