Many staffing experiments in government begin with broad goals, like reducing time to staff or improving diversity, but few set concrete targets at the start of the project against which all interventions and features are measured. As a result, few of these experiments and initiatives deliver the results they hope to achieve.
Highlighting the value of targets and measurement is one of the many valuable things the Results and Delivery work championed by the Privy Council Office set out to address. We heard them. We’ve built Results and Delivery not only into how we measure our project’s performance, but also into the design of the staffing process itself. (For more on this, take a look at the Impact-Driven Staffing design our team tested with managers.)
In the beginning, Talent Cloud set ambitious targets for reducing time to staff and improving diversity and culture fit. The first was measured in time, the second two were measured in qualitative analysis. The first was intended to deliver a new process; the second two were intended to deliver a new methodology that would generate a different outcome.
Once set, we concretely tied these performance objectives directly into the configuration and operations of the platform itself. They were not seen as a separate departmental responsibility. Instead, every form field, every operation on the platform was questioned, user tested and measured against its ability to advance these aims, while complying with all applicable government policies and requirements.
Dozens of experiments were run related to these performance objectives on all elements of a staffing process that could be influenced by Talent Cloud. This included components such as the design of posters, the HR-to-Manager communication relationship, applicant experience and online tools to help with key tasks. (Notably, this did not include elements residing exclusively with authorities outside the influence of both partner departments and Talent Cloud, such as language testing and security screening.) These experiments never met the standard for a Randomized Control Trial - we never had the volume of applicants to do this. So all the results in this report should be read as signals rather than as solutions. Nevertheless, the findings are encouraging, and are consistent with private sector and academic research findings.
When we started, it became rapidly apparent that no two HR shops in government did things exactly the same way. Some left managers largely to their own devices, while some ran a full service staffing model. Some pushed pools as the solution, while others expected managers to individually staff positions with very specific merit criteria. Some partners were under the Public Service Employment Act (PSEA), while others had increased flexibility and different processes altogether in key areas. Some departments even had “universal criteria” for every job advertised in their department, regardless of classification, and we carefully built these requirements into the default settings for specific departmental users to keep the process smooth and service-oriented.
Because of this variety (which is never going to go away) Talent Cloud was designed as a process that works for a wide variety of HR cultures in small and large departments, inside and outside the PSEA. This was the bar that was set for all features, and something we tested against rigorously. And not all tools are perfect in this regard - many are still in the process of getting upgraded to the model that came from live testing, where we did secondary testing on the things that worked for some but not for others. But in all cases, we found there was a way forward - it just took time, meticulous attention to departmental needs, and perseverance to get to a solution.
It takes all parts of the ecosystem to produce a forest. Water, seed, soil or sun alone is not enough. Only when they are working together can they grow a tree.
When departments take care of their own mandates, they provide service to citizens that is adequate. When departments collaborate, they provide service to citizens that is better than adequate. But when departments invest in the excellence of other departments, raising the quality of the ecosystem as a whole, that is when they provide service to Canadians that is truly exceptional. Because in doing so, each team, each department, works to achieve more than it could alone, and in turn is accelerated through the knowledge and growth of other teams and departments.
Service components connect as teams connect. When each department looks to the health and strength of the total ecosystem as its primary goal, it is the ecosystem that becomes high functioning and capable of greatness. And for citizens, nothing less than a high-functioning ecosystem in government will serve them well in the digital age.
So how does all this relate to Talent Cloud? Because the problems associated with staffing touch more teams across government than you can imagine. Talent Cloud was designed as an experiment to accelerate the talent engine of government, tackling problems at the heart of the ecosystem as interlocking parts of the whole. We looked at the totality of impacts, from the stress and time invested in crafting a poster through to how a hire was doing after a year on the job. We looked at manager procrastination behaviours and IT security considerations in the configuration of designs. We thought about what would waste a single mom’s time on a Friday night, and what keeps a Deputy Minister up on a Saturday night, thinking about unmet skills gaps in the organization. We thought long and deeply about who gets the job and why, and whether some of these reasons needed a rethink if departments are really committed to diversity and inclusion.
Talent Cloud’s progress against reducing time to staff and improving diversity and culture fit in hiring is a direct result of viewing technology, people, operations and policy as part of a seamless ecosystem. All intended outcomes must be successfully navigated across all barriers and throughout all components. It’s the team’s belief, based on our research and external data from other models, both domestic and international, that no meaningful progress can be made against these specific performance objectives if the staffing exercise is viewed as a relay race where the baton is passed from authority to authority, rather than as a ecosystem. If individual actions and processes do not adhere to the concrete performance target that has been agreed to, the process as a whole will not achieve it.
Notably, this is not a question of changing any authorities as defined by policy and legislation - it’s a design problem. Government authorities must do more than meet the policy requirements of their mandate. Each must look to the downstream consequences of every choice, every action, every form field added in an application or process step created. Each design component must interlock seamlessly to serve the intended outcome. It must be tested with real humans, and not be considered complete until the outcome of the testing meets requirements. User testing without changing the product based on results isn’t user testing; it’s pre-training a small handful of people before the product is released.
