The Art of the Possible

“If you’re not changing something, you’re probably holding back change.”

Public servants across government are working to improve digital service delivery, but efforts to date have not been enough. We need to put the wind at their backs with concrete, bold actions that help them deliver services that are easy-to-use, fast, inclusive, reliable, safe, and transparent. At last year’s FWD50, the Canadian Digital Service (CDS) said “challenge accepted” to the Innovation, Science and Economic Development Canada’s Digital Industries report, which implored the federal government to provide all of its services online by 2025. This year, CDS CEO Aaron Snow will show what’s possible and how we can meet those expectations.

You can see Aaron’s presentation slides here, continue the conversation with Aaron on Twitter or Linkedin.

Please see the full transcript of the talk below.

[00:00:00] Aaron Snow: Hello! Hello! Hello. Bonjour!

Okay. Je suis désolée, je ne parle pas français, je suis American. I used to work in the US government in Washington, DC and speaking of Washington, Alistair, before I begin, I just want to say to you and everyone else who came down from Montreal, congratulations on finally winning a world series.

Yeah, I’m from DC, but I’ve been here long enough to feel like I can say welcome everyone to Ottawa for FWD50 and a very Canadian sorry for the weather yesterday. You caught us somewhere between the beautiful summers that are now experienced and the beautiful winters. Okay so [00:01:00] it’s been an adjustment for a guy from the States but everyone here has been absolutely lovely and welcoming. 

I came here to help build a thing called the Canadian Digital Service. So what’s a Canadian Digital Service? Well, let’s read from one of the founding documents: the organization was created to change the way the federal government designs and delivers digital services. Just a little over two years ago and we we’ve been humming along and we were chiseling away at it, when, last fall, a report was published. The government had a digital industries table. Apparently committees are called tables up here. That’s alright. This committee had met all year. It was a dozen leaders from industries here in Canada, chaired by our neighbor just down the block, Tobi Lütke from Shopify and the report made a lot of really useful recommendations.

But one absolutely just stood out to us at CDS and it was this one. It [00:02:00] was to digitize all the public facing government services end to end by 2025, and this called out to us. It’s why CDS exists. So, last year at FWD50 2018, I stood in this very building and I said, “let’s do this, challenge accepted”.

So that’s yeah, that’s me doing that. I didn’t really give my team a heads up that I was going to do that. So that, that’s my teammate, Wendy laughing to herself, “Ha ha very funny boss. Uh, wait, what the hell did he just sign us up for?” Okay, so sorry, not sorry. I said we’d work on figuring out what it would take to meet that challenge. And we did that. 

In May, we published the beginnings of a roadmap to get us there. And maybe roadmap was the wrong word. I wonder if maybe it was more like a self help manual for government. So we [00:03:00] didn’t propose to build specific things or solve specific problems or improve specific services. The roadmap is a “how” document, a guide to improving the how we design and deliver services together. 

So what does it say? It says first to prioritize and empower and measure digital service development. And that sounds simple. But that doesn’t always happen in government. Assign, empower a single service owner. We said assemble a small multidisciplinary delivery team, give them access that they need, the systems and data and institutional knowledge and assess regularly against a set of useful standards and identify outcome criteria and collect baseline metrics and measure and publish your progress in the open.

We said second to accelerate the delivery team’s progress across the board, commit as a government to providing them with well-designed common services and components and resources that meet their development needs delivered by dedicated [00:04:00] product teams. 

We said third remove obstacles to digital service delivery. And that includes things like making cloud services easier to obtain and use, better aligning HR policy training and education and hiring and promotion and retention, ensuring access to appropriate tools and resources, enabling continuous delivery practices, replacing waterfall project gating requirements, making funding models more agile friendly.

And then last we said whatever you do, measure what’s working before planning what’s next. And this isn’t just guidance we’ve been providing for the rest of government. CDS has been working on making all of this come to life in the work that we do. So, step one, prioritize and empower and measure digital service development.

When we partner with departments to deliver services together, this is what we do. We build small multidisciplinary teams. We do everything we can to make sure they are fully empowered to deliver. And we measure the [00:05:00] outcomes as we go. So early on, we partnered with Veterans Affairs Canada to make it easier for veterans and their families to find benefits and programs and services that are relevant to them. And shout out to our Veterans Affairs colleagues who I think are here today, and also to the OG Code for Canada fellows who were part of that first team somewhere here. Alright, they’re still sleeping. 

Alright. Okay. We partnered with immigration, refugees, and citizenship to work on making the process of becoming a Canadian citizen, more intuitive and welcoming. We partnered with the center to build a platform that lets departments post challenge prizes. And we worked with Natural Resources Canada to open up access to their home energy ratings data. We worked with the office of the CIO to monitor, you know, whether federal government websites are using modern privacy and security measures, like enforcing HTPS.

And now [00:06:00] in our third year we’re working on some bigger impact projects. We’re working with CRA, the Canadian Revenue Agency, to help Canadians with low or no income, get the tax refunds and benefits they’re entitled to, which they can’t do until they file. So we’re trying to help them make it easier to file taxes.

We’re working with the RCMP – the Mounties – to make it easier for Canadians and Canadian businesses to report scams and other forms of cyber crime and help authorities investigate those crimes. Law & Order. We’re working with ESDC, the Employment and Social Development Canada – it’s always a mouthful – to make it easier and faster for people to apply for and get a disability benefits under the Canada Pension Plan.

And reality check, no lie, this is hard. For all the  reasons that many of you understand and encounter every day in government. So CDS works every day [00:07:00] at making it easier not just for us and for the partner we’re working with that day, but for everyone. Which sort of leads us straight into roadmap step two: investing in developing common services and resources.

And we are doing that. We are building the Legos that will help kickstart bringing digital services to life across the board. So here are a few examples of the several that we’ve been tackling. 

One, we made a tool to generate preapproved and compliant consent forms for folks, for government staff to use when conducting design research. Thank you for conducting design research!

 Two, we have a thing called the CDS Starter App, which makes it easy to spin up and deploy new fully policy compliant websites and services in the cloud with continuous automated security and accessibility assessments.

And three, borrowing work from our friends in the government digital service in the UK. You’ll see it above, we’ve spun up something called GC [00:08:00] Notifications, a service that makes it easy for government offices to easily and cheaply, at scale, send messages via email and text and eventually paper letters too. Go check it out. 

So roadmap step three, removing and unburdening full form obstacles that hinder digital service delivery. We’ve been hard at work on that to, bringing in the first Code for Canada fellows in government, and then helping other departments make that easier to do. Making sure that public opinion research guidance distinguishes and enables design research in government. Compensating research and testing participants for the first time, conducting assessments, making sure folks can obtain nonstandard hardware when that’s the best tool for the job and getting unfiltered wifi and so on.

So, that’s all good: steps, steps, steps. One thing at a time. We wanted to also sort of [00:09:00] test and find out, you know, what if the wind were really at the back of public servants? So imagine if the whole public service had that great set of Legos, to build services with, and no obstacles and putting delivery first. So we decided we’d just try it. We wanted to show what’s possible. 

So while we have teams hammering away at tax filing and disability benefits and cyber crime reporting, we also put a small team of three on an Island for a couple of months with a different kind of problem. And they landed on the ever beloved government form. So we pick the passport renewal form. It’s not to pick on that form or that office or that department. We didn’t ask permission. Sorry, passport folks. It could have been any form in any department. There are thousands of them. This is a learning exercise for us, about the how not the what.

So what did we do? So how does it work today? Today, you download a [00:10:00] PDF. You fill it out. You can fill it out on your computer but it’s a PDF and you gotta have a reader. And then you print it, and you’ve got to find a printer if you don’t have a printer. And you get an official photo taken, somewhere and stamped. And you pay a fee, manually, check or write your credit card info and you mail it all in. And then you wait. 

So first the obvious thing, right? So what if you could do this on the web and it were mobile responsive and accessible from the start. Okay, so the team duplicated all the same questions that are on the form, put it online. And then they made it so when you answer those questions, it populates the PDF for you. So you can print it out and mail it in. One step. It’s a place to start. What does it buy you? Well, instead of a static form, now we can use user feedback to continuously improve the content, we can provide online help, and so on. Buys us a little. 

So now, do you have to fill in all those fields every time you fill out a form? So here’s a little [00:11:00] cleverness that the team found. There’s a package, an open source package out there that makes it easy to pull data from a photo of your current passport. So the team added it. Saves everyone a few minutes and potentially some typos that can sometimes turn into real hassles when you’re talking about something like a passport. So there’s a little time and money. 

Right now, like I said, this government office expects a piece of paper mailed to them, so we can generate one for you to print out and mail in. But it’s just as easy to take that one small step into digitization to, you know, the code could email the PDF straight to the government office once they’re ready to receive that volume of emails. That’ll save everybody a little more time and a little more money and then take one more small step- if and when they’re ready to receive structured data, instead of a PDF. The code could just as easily do that too. Loose coupling. Alright. 

And hey, some other countries allow you to take a photo on your phone and even help you get that right. So [00:12:00] Canada’s policy prohibits that right now. But if it ever changes, adding a little code to upload your own photo is easy. We know because the team did that too. And even test it out some code that does a decent job of predicting whether the picture you uploaded, it’s going to comply with the requirements for the photo.

So left to their own devices, using good Lego’s, including CDS Starter App, in about eight weeks, this little team of three built something that can help a lot of people, a little. Save a little time, a little money. And all that time and all those dollars add up for people. And for Canada. And multiply that by however many forms there are out there and it turns into real money in real time, in real life. And because we’ve developed the front end independent of the back end, we keep things loosely coupled, we can continue to develop on both at the same time and independently, right? The front end doesn’t have to wait for the back end to [00:13:00] update before it can make changes that help people, that save time, that save money, that do all those things.

Someone’s going to tell you that these small incremental steps are all fine for new websites and fancy, flashy things, but they’re not going to solve the big technology thickets in government. That the really important projects need to be solved the way we always solve them. Right? You know, big budget, big plan, big requirements.

I don’t believe that. I think the evidence is clear internationally, across the private and public sectors at all levels, that the biggest risk is not to do anything differently. This isn’t easy. Governments aren’t optimized for service delivery, let alone change. But we can do better. You can’t change everything at once, much as sometimes we would like to try, but if you’re not changing something, you’re probably holding back [00:14:00] change. 

So this is my pitch to you, my solicitation: do something differently. Go home from this conference and pick one thing that you’re going to do differently. Take one small step along the roadmap. Something to help empower digital service development teams, something to invest or use or help improve one common service or remove one obstacle to delivering one great service.

If you want some examples of the one thing, I have examples  of the one thing to take home. Watch one person use your whole service. Make your teammates watch too. Then watch another. Invest in a team. Invest in a team to explore and then solve a problem instead of in a project to buy or build a particular technology solution first, or try both at the same time and see what works. Optionality is good. Redundancy is good. Redundancy is resilient. Big IT projects are nearly always risky, expensive bets. So bet on a team. 

[00:15:00] Ask why one big project shouldn’t be broken into multiple smaller parts. Pick one, pick any one. Insist on a service. If you’re involved in building or buying software in government and your team doesn’t have a service designer, insist on one. If you don’t know what a service designer is, come talk to us at 12:30 at lunch on the Blue Stage. If your team doesn’t have a product manager, not to be confused with a project manager, insist on a product manager. 

Make sure you’re measuring outcomes, not just outputs. Ask how you’re going to measure success on the project you’re working on. Ask if what you’re measuring is measuring outcomes or output. Ask whether the baseline metrics of before data have been collected. Ask if they can be. And while you’re at it, ask if you’re spending more energy managing project risks, deadlines, budgets, versus mitigating outcome risks that, you know, getting something done that helps.

Prioritize small. Ask if you’re conflating the size of the project budget with the importance of the work. [00:16:00] Big budgets, get attention, but more smaller projects are better and sometimes more important. We all know this in this room that smaller is faster. Smaller is more securable, easier to improve, easier to replace, easier to understand and explain.

Deploy “Hello World” all the way into production. First, early on. Then do something to make deployment a little faster, then make one change. If you can’t make changes continuously, much of the rest of what you do will be slowed. You will not have momentum without continuous delivery. Automate one test, then automate another. Add one API, then add another. Reconsider, one process. If it seems like there’s a policy or an in house rule, or that’s something that’s preventing you from doing any of these things, ask whether it’s really a rule. Ask for the source of the rule. Ask what it would take to update the rule. Ask whether improving service delivery might be more important than not updating the rule. [00:17:00] Learn about the strangler fig. It’ll change how you think about de-risking legacy system replacement. 

When I first joined CDS, we set objectives for the year. And objective number one was continuously improve CDS because if we don’t hold ourselves to account for changing, it’s easy to slip into being too busy to change, too resource strapped. Just keep doing what you’re doing. Keep driving in the car. That only goes 50 miles an hour instead of stopping to replace the engine with the one that can go a 100 miles an hour. 

The urgent trumps the important, right? So make it a performance goal for yourself and for everyone to do one thing differently, to make one improvement. And whatever you do decide to try differently, measure whether it helps. Learn from it and then do something else differently.

And finally, don’t learn alone in the dark. Learn together, share what you’re working on. What’s working, what didn’t work. [00:18:00] We spend an hour every two weeks at CDS on this as a team and it is an invaluable ritual. We ritualize it to make sure that it happens, to make sure that we keep doing it, that we keep prioritizing it. Because it is easy, especially when the pressure’s on, to decide that you’re going to go hide for a few months and just work on your thing all alone and not show anybody until it’s perfect. And that’s always a bad policy. 

So what does doing one thing differently have to do with meeting the digital industry table’s big, big, big challenge? I think it’s this. I think it’s Whong’s law. We’re organized and taught and encouraged at times to attack big problems with big solutions. We know big bang IT projects fail more often than they succeed, but Whong’s law persists.

So enough. Enough. We know that usable, adaptable, resilient systems don’t emerge whole from a grand plan. They [00:19:00] emerge, they evolve as we build and measure and learn from them. They are the cumulative result of small changes, of doing something differently, and then something else, and then something else.

So if we want to digitize all public facing federal services by 2025, this is how it’s going to get done. This is how we’re going to put the wind at the back of civil servants. So are you in? Are you in? Are you in? Yes, alright! Good, alright good! And if you don’t know what to do differently, you don’t know where to start, you need help trying, come talk to us. That’s what we’re here for. That’s what CDS is here for. To do everything we can to put the wind at your backs. So come have lunch with us today at 12:30 on the Blue Stage. Merci beaucoup, merci beaucoup, merci beaucoup, merci beaucoup. Thank you for having me. [00:20:00]