Making the Canadian Digital Service Bolder
Acknowledgements
I am thankful for conversations with Ryan Androsoff, Co-Founder of the Canadian Digital Service (CDS) and current independent digital government consultant, with Olivia Neal, formerly from the UK’s Government Digital Service (GDS) and now Executive Director, Digital Change at the Treasury Board of Canada Secretariat (TBS), as well as with Sean Boots, Technical Advisor at CDS, and Annie Leblond, Communications Officer, also at CDS. Their insights were invaluable in understanding CDS and their passion to make government better for citizens is remarkable and inspiring. This work has been done in the context of a course called Disrupting Bureaucracy: Understanding Recent Efforts in Digital Government taught by David Eaves at the Harvard Kennedy School.
All errors or omissions in the following work are my own.
CDS has a promising foundation
The Canadian Digital Service (CDS) was officially founded in April of last year, with a funding allocation from the 2017 Budget of about $25 million (CAD) over three years. CDS was not assembled to respond to a major IT project disaster like the United States Digital Service (USDS) with HealthCare.gov nor was it born under strong fiscal pressures for cost savings like the Government Digital Service (GDS) in the United Kingdom. Rather, the organization was created as a result of an organic evolution of innovative work done by bureaucrats within the government as well as a deliberate public consultation process across Canada.
The “Beginning the Conversation… A Made-in-Canada Approach to Digital Government” report is the product of conversations with citizens from across the nation about their vision for digital government and includes perspectives from the private, public, and social sectors. It was published on July 18, 2017. One of the key insights from the report was that “People expect government services to be simple and easy to use” (now the official tagline of CDS). In essence, the public sector needs to step up and provide digital services that are as easy to use as those provided by the private sector. Makes sense. To achieve this, the report concludes that the government of Canada would need to embrace modern software development tools and approaches, plug into the talent pool of the private sector, and create a digital service group like those created in the UK, USA, Australia, and Ontario to lead a digital transformation within the federal government.
CDS is in a good position. It has broad-based support within Canadian civil society. It secured three years of no-strings-attached funding (the first of which has run out), and also has strong political support. The Honourable Scott Brison, President of the Treasury Board, is a strong sponsor and believer in CDS. He gives the organization space to innovate. Finally, CDS is deliberately housed in the Treasury Board of Canada Secretariat (TBS) — the center of rule-making in the Canadian government. For example, this is where the federal CIO and his team set IT policy for all of government. CDS does not report to the CIO, but its creators hoped that this organizational design would help the two groups be aligned. This seems to be the case.
Additionally, CDS’ privileged access to policy-makers has given it the ability to clarify rules and ensure that ‘de facto’ interpretations of existing policies were not constraining delivery unintentionally. For example, CDS has been able to use various HR innovations to attract top digital talent. At this point about 60 high-capability staff members have been recruited to the new organization. CDS’ link to TBS also gives it legitimacy to try new things when working with government partners across the government. Overall, CDS is in a very strong position to effect change across the Canadian bureaucracy.
Embracing citizen-centric and Agile methodologies
Modern, Agile software methodologies are at the core of the delivery model of CDS. For example, the project teams follow an iterative design approach and publish alpha and beta prototypes online as soon as possible. According to blog posts from Jason White, the Head of Data Science at CDS, the team leverages significant automated testing and aims for continuous integration of contribution to their codebase. Basically, they work like a modern software shop — this is great news.
Figure 1: CDS releases software iteratively, following modern best practices
As explained by Sean Boots, Technical Advisor at CDS, teams are also multi-disciplinary and typically include 1–2 developers, 1–2 designers, and 1 product manager. Each team also has support from a policy expert that understands government and a communication officer to tell the team’s story both within and outside government. These folks tend to spend about one third of their time per project team. Working with end-users is a core part of CDS’ approach — teams often co-locate within government agencies and work closely with developers from across the federal bureaucracy.
Similar to other government digital units across the world, CDS appears to have a ‘delivery-first’ attitude. CDS aims to leverage open-source — both by integrating existing solutions in its products as well as publishing its code on GitHub. Working in the open has several advantages — the latest technologies can be leveraged, more feedback can be incorporated to reduce bugs, and CDS’ solutions can create additional value by being reused both inside and outside government.
Traditionally, the government (not unlike big Canadian banks, or other Fortune 500 companies) has been using Waterfall design methodologies and works on long release cycles. It has also tended to work with proprietary technology solutions and did not share its tech journey externally. This is changing. It is refreshing to see CDS’ new approach to delivery.
But, CDS has limited responsibilities compared to peers
The Canadian Digital Service (CDS) has a generous and predictable budget enabling it to offer its services to government agencies at no cost to them. It has strong political support. Its newness and strong branding is creating excitement and enabling the organization to hire top digital talent and attract partners within government.
This might not last. Funding must be renewed in Budget 2020 — in less than two years. The Liberal government is up for re-election in October 2019. Political support is therefore less predictable in the future. Finally, excitement and goodwill are not a long-term recruitment strategy for both internal partners and external talent.
CDS seems to be avoiding taking risks, limiting its scope for transformative impact. Based on my conversations, avoiding thorny projects has been a deliberate approach. This way, CDS can focus on delivering value without rustling feathers and safely gain a positive reputation. CDS does not have the mandate to consolidate digital services across the government like GDS did with GOV.UK. Big platform projects like identity management across jurisdictions is outside the scope of CDS and is instead being pursued by the CIO at TBS. While procurement reform for IT has been a major savings lever in the USA and the UK, this is also outside of CDS’ scope in Canada. Finally, CDS has avoided getting involved in major IT disasters like Phoenix Pay or Shared Services like USDS did with HealthCare.gov.
Figure 2: CDS has limited scope for transformational impact compared to peers
Don’t get me wrong — this humble approach has merits. It has been argued by David Eaves and others that GDS potential for impact has been severely limited within the UK government by backlash from the rest of the bureaucracy due to its cavalier approach. As a new organization, CDS can benefit from these lessons and move forward with a more collaborative approach. Nonetheless, CDS should be bolder. Canadians deserve high-quality, customer-centric digital services. Additionally, political and financial coverage will not last forever. Being cautious could be more dangerous than having a clear theory of change and using the next few years to test and refine it aggressively.
To create real change, CDS must articulate its strategy and test it ASAP
Digital service groups across the world have not converged to one, unique theory of change. Digital transformations are not a solved problem. Context seems to matter, a lot. CDS is a start-up. Like all start-ups, it must find its product-market fit.
Digital service groups have particular constraints. In general, they are much smaller than the bureaucracies that they want to transform. For example, CDS is planned to have about 60 to 80 full-time staff at run-rate. On the other hand, the IT function in the federal government employs tens of thousands of government workers. It is clear that if CDS’ goal is to truly be transformative, it must find high leverage strategies to help government evolve.
Figure 3: CDS needs to articulate a clear theory of change
Based on approaches from existing digital service groups, there seems to be four promising transformational levers in digital government. CDS could: 1) Develop digital skills across government, 2) Build government as a platform, 3) Reform IT procurement, or 4) Rescue major IT disasters.
Develop digital skills across government
One theory of change for CDS would be to focus on scaling digital approaches across the government. This would require identifying current staff with digital skills, training those who don’t, and making changes to recruitment to ensure that top digital talent can be hired within the federal Canadian government. Compensation, working culture, career progression, and flexibility to work both within and outside of the government are potential levers to improve recruitment of digital talent. Based on its learnings, CDS should be directly involve in changing HR policies.
In say, five to ten years, the goal would be to either have digital groups across all government agencies, or even better, make Agile, user-centric development practices the default way to develop software in the government. As an aspiration, policymakers should also adopt Agile practices (e.g. collaborating with end-users, iteratively testing programs, and being ruthlessly data-driven) when doing their work.
CDS is creating demonstration projects that can be used to show the art of the possible when a team uses modern practices. While this is not enough to be transformative on its own, it is an important and necessary part of the journey. It can serve as ammunition for staff wanting to do things differently. It can inspire other who previously didn’t.
CDS’ use of co-location on projects is a direct way to diffuse best practices across the organization. Unfortunately, this is probably too slow to fully transform a bureaucracy as large as the federal government. To accelerate this, organizations like GDS have held trainings across the government to share new practices with government staff. This could be a tool for CDS to experiment with over the next few months. In each agency in which CDS works, the goal could also be to leave a ‘digital hub’ assembled from bureaucrats from the client agency. Having this as a clear metric for the team would make sure that training and upskilling is a core part of the group’s agenda.
In the private sector, digital transformations are often accompanied by large layoffs and subsequent hiring of staff possessing the missing skills. This is not a feasible option — nor a goal — for the federal government. On the other hand, labor inflexibility limits the ability of the organization to adjust quickly to technological changes. This is a constraint that must be kept in mind if CDS chooses to explore this lever further.
Build government as a platform
Another approach to scale scarce digital talent from a digital service unit is to empower this team to set overall design standards for the government and work on platform technologies. By developing — and enforcing usage of — the most crucial pieces of technology, CDS could have an outsized impact on services delivered to citizens.
This approach was used by GDS in the United Kingdom. All government webpages were transitioned to GOV.UK, a one-stop web portal for citizens controlled by GDS. Additionally, GDS focused on core technology pieces like GOV.UK Verify for identity management that created a superior digital experience for citizens across all government services. Similarly, in Estonia, digital staff has innovated on the architecture of digital government with positive results. For example, the organization has built X-ROAD, a shared data platform that can be leveraged by both public and private platform to securely access citizen data, as permitted.
CDS currently aims to choose projects based on three criteria: 1) Readiness, 2) Reach, and 3) Replicability. Readiness means that a government client already has a well-defined problem statement and is willing to work following Agile methodologies. Reach means that the solution will either impact a lot of users or make a big difference for a smaller number of users. Finally, Replicability is the closest criteria to working on platform innovations — solutions should be scalable to other users or part of government.
In practice, in its first year of existence, GDC seems to have been opportunistic and tackled projects driven mostly by their readiness and the capacity of their organization (at the expense of reach or replicability). For example, under ‘Past Products’, the GDS has created two products. The first project is an alpha of an E-Briefing app to replace binders (most likely not transformative, but a good demo of the team’s capabilities). The second project is a beta of an Impact Canada Challenge Platform (again, most likely with limited reach and replicability).
There are also three in-flight projects. The first one is an alpha of EnerGuide API — an API to “open up access to EnerGuide Home Energy Ratings” in a reusable way. Experimenting with APIs could have replicability, but there is most likely limited reach for this particular product. It is not a game-changer.
The next two projects are a bit more ambitious. There is an alpha for Rescheduling Service for a Citizenship Test. This can have an important impact on future Canadians and code for rescheduling appointments has the possibility to be re-used elsewhere in the organization. Finally, there is an alpha for an Online Tool to Improve Veterans’ Access to Benefits. This is also meatier and ranks higher on reach while potentially providing modules that could be reused elsewhere in government.
Based on my interviews, other foundational pieces like cross-jurisdictional identity management are under the purview of the CIO within TBS and there is no current plan for other transformational projects like Estonia’s X-ROAD in Canada.
The main risk with this approach is that it can antagonize the rest of the organization. As discussed earlier, this is most likely one of the key reasons why GDS’ progress has recently slowed down in the United Kingdom.
Reform IT procurement
The Government of Canada spends hundreds of millions in IT costs each year. This spending tends to be focused on a few large incumbent firms. Limited competition and an outdated Request for Proposal approach designed for Waterfall software projects is costing the government too much. Reforming procurement to enable more small and medium-sized players to serve the government is an approach that has been taken in the UK, USA, and Estonia to transform delivery of digital service. Procurement reforms also introduce a focus on outcomes instead of outputs, to better align the incentive of vendors with government buyers. Finally, having public servants that write high-quality software within the government on procurement teams tend to improve the quality of procured goods and services.
In Canada, reforming IT procurement is outside of the scope of CDS. Instead TBS is working directly with Public Services and Procurement Canada (PSPC) to iterate on a new approach called Agile Procurement. This work looks promising and was initially piloted based on a problem statement to fix the Open by Default Portal of the Canadian government. One potentially transformational lever for CDS would be to plug into this work and help the teams improve their approaches based on their technical expertise. Another way to influence procurement would be to make CDS sign-off necessary on new IT procurements and projects. While again this carries the risk of alienating other bureaucrats, it is a lever that can lead to large impact with a small team.
Rescue major IT disasters
Canada has had its share of disasters in IT procurement in recent years. For example, Phoenix Pay — a payment system for federal employees — has cost more than a billion dollars and does not work. Shared Services, an attempt to centralize several back-office functions has also been a disaster. Rescuing disastrous project would be a role with possibility for high impact for CDS. This was the model used by USDS when they fixed HealthCare.gov.
Based on my discussions, this does not seem to be the best use of CDS talent and energy at this time. The skills of the existing team are not necessarily compatible with this type of work and solving these major projects would not directly improve the experience of citizens with the Government of Canada.
CDS should double-down on developing digital skills across the government
According to CDS’ website, its current strategy is based on delivering solutions, building capacity, and providing advice. Based on my interviews, CDS so far seems most concerned with delivering solutions. This makes sense — having a few successful projects would increase the credibility of the group and enable it to make more changes across government.
Based on how CDS sees itself, focusing on developing digital capabilities across government seems to be the transformational lever closest to its current identity. My belief is that CDS should fully embrace this role and double-down in making it happen.
To test the hypothesis that this is a good strategy, I have three recommendations: 1) When CDS leaves a department or an agency (e.g. Veterans Affairs Canada), it should have an explicit goal of leaving behind a digital hub. Members of this hub should be officially affiliated with CDS through a community of practice (e.g. data scientists, user research designers, etc.) and have access to continuous trainings. The goal would be to have ambassadors for digital practices across the bureaucracy. These professionals should also have the opportunity to work in other departments on a short-term basis to further cross-pollinate the Canadian Government with digital mindsets and skills. 2) CDS should staff a team to pilot Agile boot camps across the government to upskill the labor force and share digital approaches. This should be tested in the next few months. Education talent should be hired if required for success. 3) One of CDS’ project should be to work with TBS to evolve the recruitment process for digital talent. Policy bottlenecks should be addressed (e.g. salaries, career planning, remote work, etc.) and technical tools should be designed to reduce frictions in the application process.
Figure 4: Recommendations to build digital skills across government
CDS is being wise by not trying to affect change through all four transformational levers — this would spread the organization too thinly and potentially create pushback from the existing bureaucracy. Most likely, CDS’ humbler, more collaborative approach is more sustainable in the long-run. On the other hand, this approach isn’t an excuse to forgo actively testing a theory of change that can create the results that it seeks, taking into account the constraints of its small size. The team should therefore re-orient some of its existing resources to testing a particular theory of change. Its current political cover should be enough to enable it to do this — it should be used before it’s too late.
To be successful, CDS should track progress against its theory of change and adjust accordingly
This theory of change is built on a series of assumptions. This is why it is important to test it as soon as possible and iterate as we learn more. The ultimate mission of CDS should be to transform government — testing an approach to do this should therefore be part of all work that it undertakes.
Clearly, CDS should continue to track project-level metrics to ensure that its software products are meeting their users’ needs (e.g. usage, retention, satisfaction). On top of these metrics, I recommend that CDS tracks three additional program-level metrics: 1) The satisfaction of current CDS team members, 2) The number of digital hubs and digital hub participants in each client agency and department, and 3) The general perception, knowledge, and usage of modern digital practices (e.g. Agile development, user-centric design, etc.) across the civil service.
Satisfaction of CDS team members
For the Government of Canada to be successful in its digital transformation, it must be able to create work that brings satisfaction to digital talent. As CDS grows, it should make sure to take its own pulse on a weekly or bi-weekly basis to ensure that it is building a positive, exciting, and inclusive culture. Being able to course-correct fast will ensure that CDS team members are ambassadors about a new way to do things both inside and outside government. Important lessons in creating an environment where digital talent thrives can also be gained from this exercise.
Growth of digital hubs
Simple metrics are the best metrics. CDS should track the percentage of projects that lead to the creation of digital hubs in their original agencies. The total number of digital hubs and participants should also be measured as well as their engagement with CDS. Satisfaction of these users should also be monitored to ensure that they feel supported as digital ambassadors.
General digital survey
Every year, CDS should develop a representative survey for all government bureaucrats to get an overview of the current perception, knowledge, and usage of digital practices. Special interest should be paid to understanding the view of CIOs of the digital transformation. Ambitious targets should be set in increasing comfort-level with digital practices. These metrics would also be a leading indicator of any organizational resistance to the digital transformation. Particular pockets of excellence or bottlenecks could also be targeted for projects by CDS to drive the desired transformation.
As a Canadian citizen, I am thrilled about the potential of the Canadian Digital Service to transform the government-citizen interactions in our country. I can’t wait to see everything the team accomplishes over the next months and years.
