Public service heroes: Chris Allison

In the federal public service, so many of the building blocks of digital government and tech modernization are thanks to Chris Allison. An early leader of the GCtools team and the CSPS Digital Academy, he’s now the Chief Data Officer at the Public Health Agency of Canada. He’s also perhaps the only senior federal public service leader who is fluent in Python programming. He’s a lifelong hero and inspiration of mine; we spoke on May 3.

Sean: For introductions, tell us who you are and where you work these days!

Chris: Hey! Chris Allison, I’m the Chief Data Officer and DG for data management, innovation, and analysis at the Public Health Agency of Canada.

Nice! I’m guessing that keeps you busy…!

It does! It does. The School was an amazing place to learn and grow, but PHAC is an exciting and very busy place to be.

I can imagine. As a lot of folks know – you were doing digital government since the earliest days, almost kind of before it was cool. How did you originally get started in the public service?

Oh – in the public service? I was living in Japan, in the year 2000, and I had been teaching there – teaching English as a Second Language in Osaka for a few years – and decided I should actually grow up at some point, so started looking around for jobs back in Canada. I applied to a Post-Secondary Recruitment campaign, ended up getting accepted for a couple of jobs, and started working with Immigration Canada as an Immigration Enforcement Officer in Toronto.

Whoa, no way. I did not know that. How did you eventually make the switch from what sounds like more front-line work to government tech?

It was a very gradual switch. I did three and a half years of front-line immigration enforcement, investigations, removals, war crimes work with CIC (Citizenship and Immigration Canada, now IRCC). I then quit and joined the Mounties. I was a Mountie for close to 4 years; then I quit the Mounties and joined the public service proper, and joined CBSA (Canada Border Services Agency).

And around that – a couple years later – I was a manager in CBSA, and I started doing a lot of work with Excel actually. And just kind of, really getting into first of all the functions and the way you could set up and automate things in Excel, and then into the macros you could use, and using it basically to do rudimentary data analysis, forecasting, things like that – that were super useful to the team and then to the organization. And that kind of started me on that pathway.

But then it was a few years later – I had joined Public Safety – I did a bunch of data analytics in Excel on HR and the PSES (Public Service Employee Survey) and then came back to CBSA, and I was responsible for a team that was supposed to build the first border app.

And I ended up in this labyrinthine Catch-22, where I was trying to work with the CIO organization to find out, “how much will it cost to build an application?” And their answer was, “well, we can’t tell you until you reach Gate 1”. Okay well, “How do I do that?” “You can’t reach Gate 1 until you have got approved funding for the work”.

So then I went to Finance and I said, “Okay, I need to have approved funding so I can get to Gate 1”, and they said, “how much money do you need?” And I said: “I don’t know. They won’t tell me.”

So I spent six months in this “Bermuda Triangle” trying to figure out, how on Earth can I get someone to tell me something about how to use technology in government? And nobody could or would.

And so I basically said, “screw it. I’ll need to figure this out myself.” At this point in time – I was playing too much MechWarrior Online – and I said, “I could probably be doing something more productive here” so I started learning how to program Python.

Wow, nice.

And that was the start of the path, I guess, to actually getting into dev and data and analytics proper.

Amazing, that’s phenomenal. That’s a cool origin story, I mean especially as one of the biggest champions if not the biggest champion of data science work in the government nowadays.

And maybe something important is, I was probably forty at the time, maybe forty-one? And no formal development training since high school when I was doing Java or something terrible. I was blown away by the ease of access and figuring things out – and then getting into projects, things that I wanted to do, and learning through those. The experience was phenomenal.

You’ve played a huge role in digital government work, from leading the GCtools team, launching the Digital Academy at CSPS, championing open source and modern tech in government from the earliest days… Have you seen a shift, from those early conversations with the CIO and finance people to, is technology work recognized at a public service-wide level? Or is it still for the most part, small teams, like “pockets of awesomeness”? Have you seen larger-scale changes over that time?

There are definitely larger-scale changes, in the acceptance and openness of government, that help, but are kind of peripheral to it in a way as well. We still haven’t dug ourselves out of the hole that probably occurred back when I was starting in government, when we thought we could outsource technology because it was this “thing” and the government would just buy it.

My feeling is, for the past twenty years, we have not had the capacity to do the technology and data work that we need to do in government, and so we’re always struggling. And when you throw a good solid dose of bureaucracy on top of that, it gets really hard to get things done.

And so you end up with a few people who are kind of the “bureaucracy hackers” – to pull on some of Nick Charney’s work – who also get technology, who are able to show up in a place and do things. And then they wow the people around them.

I’ve seen more and more of these places gaining traction, and doing amazing work once those skills are recognized, but it can still be really tricky depending on where you are.

That makes a lot of sense.

One thing I really like from your work is how tangible it is – a lot of efforts at digital government are often sort of, high-level like “here’s a giant strategy!” Versus, you have a list of awesome tips for data and AI work in government and it’s like: here’s the things you need, and it’s super down to earth. Do you find that that’s sort of a challenging space where people are at different levels on that?

No, I think it’s – definitely not challenging, I think it’s been super useful for my work and career. In terms of, thinking about value – strategies are great; you need to have one. But how much time do you need to build it? What does excellent look like? What are the opportunity costs of all the governance and planning work we do? What else could we achieve if we hit “good enough” and then got to work?

To frame this, practically, PHAC’s data strategy is really good. It was put together in 2019. It hits all of the things you need. You need interoperable systems. You need talent. You need people. You need access to good data. You need the tools and platforms so people can use the data.

And so coming in, even two and a half years later, we were able to look at it and say – you know something? This is good enough. It’s got all the basics that we need. We don’t need to spend another year and half reimagining the strategy. Let’s focus on executing it.

That’s awesome.

Bit of a jump but – if you had to choose just one moment in your public service career that was extra memorable or extra inspiring, what would that be?

That’s a great question. It would probably be the graduation of the first cohort from the Canada School of Public Service’s Digital Academy, the Digital Academy Premium.

Just seeing what public servants did, right? It was a hypothesis, we didn’t know if it was going to be true. But: what happens when you give public servants the tools they need? We gave them dev laptops, we trained them on Python, trained them on R, trained them on DevOps, trained them on design… We gave them time to do this stuff, without all of the constraints and other time sinks that we have in government, and gave them each other. We built these small multidisciplinary teams – taught them agile practically over the course of two weeks – and they absolutely rocked it. Fifteen teams.

This is what people can do. Folks in government. We’ve got the people we need; we need to give them the tools and the skills and invest in them, and just get the hell out of their way. Let them do amazing things.

Seeing that was awesome.

I really like that. It lines up really well with my own philosophy about how we can make the public service more awesome.

This is sort of a classic “Ask Me Anything” question that we have at CDS. If you could wave a magic wand and change one thing about the government or the public service, what would that be?

I ask that question too. But I don’t think it’s been asked to me in a long, long time.

I would crank up our tolerance to risk. And I’d actually change it to an appetite for risk. And not in a stupid way, but in a recognition that you don’t get to success in one hop. That it’s an iterative process.

And that everything we do – like, life is beta. Everything that we’re doing is, “hey, this is the thing that we’re going to do, and we’re going to do it, and we’re going to learn from it, and guess what: it’s gonna suck, because it’s the first thing we’ve done. But the second thing that we’ve done is going to come two weeks after that, and the next thing is going to come two weeks after that.”

And so we get out of the analysis paralysis space; we’re able to cross the execution gap because we’re not afraid of doing things. And I think, with that attitude, the people we’ve got – their ability to acquire those skills and do the things, and continue learning, based off that circuit – will really give us the workforce, the culture, and the proven track record of success, that government needs to have. And that our people in Canada deserve.

That’s a great way of putting it. I remember in an early CDS partnership, I think we were launching a website with a partner team, and there was lots of hubbub around “is the content perfect, is it finalized?” And I think it was Ryan Androsoff was like: “we can change it, after we launch”, and there was this jaws-dropping kind of moment. Of like, that’s a real change: things don’t have to be perfect the first time around, you can change them.

So I was actually with Ryan back in the GCtools, we had an overlap of four months when he was on his way out and I had come in. There was a moment in that point in time – and it’s kind of funny, because the GCtools team became a model for the multi-disciplinary team –

In a huge way, yeah.

But at that point in time, there was literally: one side was the business team, and there was a long empty hallway with nothing inside, just a long empty hallway, and the dev team was on the other side.

And the way that the team worked was, the business team would draft an MOU every year with “these are the deliverables that we need to have from the dev team”. And the dev team would update it, and then both directors would sign it, and then, that was basically it.

So the business team would be sitting there waiting for the dev team to do their stuff that was in the MOU.

Whoa, I did not know that!

So, one day there was a bug on the platform. The text in the search bar had turned to white for some reason. So you couldn’t see what you were typing, because of course the background colour is also white, so you’ve got white text and no one could see what the hell they were typing.

So it was like: oh my gosh, what are we going to do? Normally, we’d submit a ticket, but that takes time, it’s a process and our users were suffering here… So Ryan and I both walked down this long, straight, empty hallway, got to the other side, went around the corner, and said, “hey look, users are reporting this”.

And one of the devs, listened to us, said “oh, okay!” and they went in, tappity-tap, changed the font colour, and it was done.

It was like, “hold on, we can do that?” Yeah, you can do that. When you have the technology people, the devs, the data people, working with the business, you can do that. You can make the changes you need, you can iterate, and you can actually provide really, really good service.

Nice, that’s awesome.

Thanks for taking the time to chat – any last closing thoughts or closing ideas? Thoughts about the future of the public service?

Yeah! That’s where I’m going. The public service needs people with the skills for the future of government. A lot of those skills are technical. Honey has a great quote: “it’s not about the technology, until suddenly, it absolutely is”. Right?

It’s about the people, it’s about the culture, it’s about the business, it’s about the design, it’s about the service – what are you trying to do? But then, I mentioned the execution gap. We need to have the folks that can help us cross that gap.

Without the engineers and technologists, without the people who can actually do things with technology, government keeps getting to the point where we’ve got great ideas, we’ve thought about it, we’ve consulted, and we hit a point where we’ve basically got two options. One, we can do procurement, which will a) crush our souls and b) take two years; or, two, we can ask the CIO, who has a dozen priorities and a thousand problems. Our brilliant idea isn’t going to be on that priority list.

And so great ideas and better services perish.

We need to recognize that data and technology are core business to the Government of Canada. We can’t actually be successful without them. We need to start building skills and practices that support their use into our organizations, build it into the DNA of the orgs, and how we operate and how we hire and how we think. That’s what a modern government looks like, and that’s where, hopefully, we’re on the path to get to.