Monday is election day – don’t forget to vote if you haven’t already!
Back in December 2019, I wrote a set of suggestions for the next GC Chief Information Officer. In the same tradition, below are some suggestions for the next Minister of Digital Government. This was a new ministerial role in the most recent government, created in 2018 as an additional role for the President of the Treasury Board, and established as a separate minister in 2019.
At present, the Minister of Digital Government oversees three organizations in the federal government: Shared Services Canada (or SSC, which runs IT infrastructure across government departments), the Office of the GC CIO (or OCIO, which sets government-wide IT policies), and the Canadian Digital Service (or CDS, which helps departments deliver better public-facing services). Full disclosure, I work for one of these organizations. 😅
Digital government work – and public service reform, more broadly, which is what it ultimately is – isn’t really a newsworthy election topic. It’s near and dear to my heart, though, and I’d love to see more conversations about it from public servants, politicians, and the public alike.
With that in mind, here are some suggestions! These are intentionally far from my day-to-day work, because, y’know, otherwise it’d be weird:
Is this transition advice or what? https://t.co/k5F8HllPmG— Michael Karlin (is on vacation) (@supergovernance) August 9, 2021
One of the challenges of the Minister of Digital Government portfolio is that – more than most other Cabinet roles – it goes deep into the “how” of how the public service works, not just the “what” of what public servants are directed to work on. That’s not something ministers are accustomed to thinking about, and it’s also not something that senior public servants are expecting to hear ministerial opinions on. The more specific these kinds of recommendations get, the harder they can be to convince public servants to implement them:
Access to information
Increase the retention period for completed Access to Information summaries. Right now, summaries of past Access to Information requests are published on the Open Government website, run by OCIO. The published list only includes the last two years’ worth of entries, which is very short for what would otherwise be a useful historical record. Changing this to a 10+ year retention period would be very little effort, and be valuable for public administration researchers and journalists everywhere. The next step could be to work with Library and Archives Canada to archive summaries indefinitely, as a publicly-accessible dataset.
Begin rolling out a “release to one, release to all program” pilot. Several U.S. states and other jurisdictions around the world have a standard program of publishing the actual substance of completed access to information (or freedom of information) requests – the documents themselves – rather than just summaries of what was asked for. This is often called “release to one, release to all”. In the past, concerns about Official Languages Act compliance and web accessibility have been raised as a barrier to doing this in Canada. Publishing released documents under the authority of Library and Archives Canada – or investing in automated ways of extracting the text of released documents, to make them more accessible – could be potential ways to address this.
Update the Copyright Act to make Government of Canada productions public domain by default. The United States, again, is decades ahead on this – any U.S. federal government publication belongs to the public domain, by default. This makes it possible to reuse U.S. government information in a huge range of other contexts, everything from analyzing climate or finance data to publishing photos of astronomical observations or military aircraft on Wikipedia. Canada doing this would be a huge boost to the open data community, to academic researchers, and to educators. All of these groups currently need to navigate the murkiness of Crown copyright to stay on the right side of the law.
- Invest in new content management system options for departments not on Canada.ca. Canada.ca was originally planned as the one content management system to rule them all. Only a few dozen departments and agencies ultimately migrated to it, however. Since the migration effort wound down, it hasn’t been clear if departments are allowed to invest in other modern content management options. As a result, many departmental websites still use incredibly old, fragile systems (or hand-crafted HTML) that limit how quickly and effectively departments can communicate with the public.
Publish department-by-department statistics on the number of in-house versus contractor/consultant IT staff. A large portion of federal government IT work is undertaken by contractors. This has a lot of implications for institutional memory, accountability, opportunities for SMEs, and vendor lock-in. Consultants have been described as “ubiquitous as wallpaper and about as unexamined”; collecting more data on how large a role they play would be a great start.
Encourage departments to withdraw requirements that CS staff only work in CIO divisions. Earlier in 2021, TBS’s Office of the Chief Human Resources Officer published a useful bulletin to departmental heads of HR, confirming that teams can include a range of classifications, and that reporting structures can include a mix of classifications. In practice, departments usually require homogenous reporting structures, and many have requirements that CS (computer systems) staff can only be hired within CIO divisions. Both of these are barriers to the kinds of multidisciplinary teams that do the best work. They also reinforce a lack of non-management software developer career progression tracks that are common in modern technology companies.
Promote multidisciplinary teams and publish reusable job descriptions for design researchers, product managers, and other design and technology roles. The United Kingdom launched the “Digital, data, and technology” career framework in 2017, in order to more effectively hire and retain digital professionals. Modern digital roles like design researchers, product managers, interaction designers, and data engineers don’t fit clearly into the Canadian government’s existing landscape of job classifications.
Large project management
Implement a maximum duration and financial cap on IT projects. This is repeating myself a bit, since this was also on my list of suggestions for the GC CIO. Large project sizes – in time, financial cost, or both – are far and away the most likely to fail. One of the most effective interventions a future Minister could make is to set and enforce a hard cap on the maximum size of projects. Anything larger would need to be split into smaller projects, that are shipped and deliver tangible value to the public, before follow-up projects are started. Actually stopping expensive, in-flight projects takes a lot of guts and political capital, though. The United Kingdom’s spend controls are estimated to have saved £1.3 billion over five years (about $2.3 billion CAD); it’s definitely a strategy worth emulating.
Require that medium to large IT projects publish public proposals, gather feedback, and incorporate that feedback into subsequent funding submissions. If an effective size cap (the item above) was in place, this wouldn’t necessarily be needed. Thus far, though, I haven’t seen any of the oversight processes in place actually stop ill-fated projects. Subjecting them all to public scrutiny would be a fascinating way to see which ones hold up. Having a department put out a plan for a major IT project for the public to see – while it’s still in the early planning stages – would be a great litmus test. Is it built with technology that typical tech companies would be embarrassed to still use? It might not be starting off on the right track.
Reduce the number of oversight committees within departments and at a GC-wide level, to approach single ownership of project outcomes. Each time there’s a major government IT failure, new oversight processes and committees are created to add more scrutiny to future projects. Counter-intuitively, though, this actually dilutes any sense of clear ownership over projects’ outcomes (and adds a lot of unhelpful administrative burden). If a project made it through 12 committees, then failed, was it really anyone’s fault? One of the main takeaways from the Phoenix review is that each IT project should have a single, accountable owner. To get there, one important starting point would be to reduce the number of oversight committees (both centrally and within departments) as much as possible.
These are just a few ideas that a future Minister of Digital Government could champion, regardless of political orientation. I’d love to hear what others think – what would you like to see the next Minister of Digital Government take on?
If you see other blog posts or publications asking this – or if you write one! – let me know and I’ll add a link to the bottom of this post.
To every Canadian citizen: happy voting! And to my fellow public servants: good luck as you finalize transition advice for the weeks and months ahead! Rock on.