Measuring the Performance of Public Services
Public services exist to serve a purpose—supporting society, helping businesses, and contributing to the economy. But how can we tell if these services are truly working well?
In large and complex organizations, different teams often focus on different priorities and measure success in different ways:
- Financial teams track transaction volumes and cost to serve.
- Operations teams focus on productivity, first-time resolution, and demand patterns.
- Digital teams measure adoption, accessibility, speed, reliability, security, usability, and reuse.
- User-centred teams assess how easily people can complete tasks, who is excluded, and the overall experience.
Why this matters
It's not just about what we do, but why we do it. Too often, we don't measure whether a service achieves its intended outcomes. As a result, we risk spending time and money "doing the wrong things well."
An end-to-end service measurement approach helps address this gap. It:
- provides a clear view of how the entire service is performing
- aligns teams, projects, and technologies around shared goals
- defines what "good" looks like in context
- supports better, more consistent decision-making
Making the most of limited public resources has always been important, but it's even more critical today.
New AI tools are accelerating how quickly changes can be made. At the same time, the environment around services continues to evolve, shaped by rising customer expectations, cyber security risks, workforce changes, and industry dynamics.
To keep up, organizations need to anticipate and respond quickly. Measuring end-to-end service performance is key to doing that effectively.
What is service performance?
Service performance tells you what "good" looks like, as a whole and across the different stages of service delivery from end to end.
At its core, service performance comes down to two things: effectiveness and efficiency.
Effectiveness comes first. It asks:
- Does the service deliver the right outcome?
- Are people getting what they are entitled to?
- Is the service achieving the intent behind the policy?
- How often does the service lead to a successful result?
Efficiency looks at how that outcome is delivered.
- How much time, cost, and effort is involved?
- Where is there friction in the process?
- How much error, rework, or avoidable demand exists?
In public services, effectiveness usually matters more than efficiency. It's often worth accepting some added time or cost if it means more people can successfully use the service. A service that doesn't deliver the right outcome can never truly be efficient. It's simply "doing the wrong thing well."
User experience is closely tied to both factors. People tend to feel better about a service when they can achieve the right outcome with minimal effort. However, measuring satisfaction alone doesn't tell the full story. It shows how people feel, but not why they feel that way, and it's that "why" that helps us understand what needs to improve.
Measuring service performance is not about setting unrealistic expectations. It's about seeing the full picture so organizations can make better decisions, and know where to invest, what to improve, and how to make the best use of limited resources.
Example
A government program wants to make its services accessible to as many eligible people as possible. In reality, many Canadians face barriers like:
- limited access to digital tools
- language or accessibility challenges
- complex or non-standard life situations
- difficulties completing applications without help
If the program relied only on a fully digital, highly efficient model, many eligible people would be left out. This would undermine the policy's goal of providing income support.
To address this, the program intentionally keeps less efficient options available, like:
- in-person support
- phone assistance
- paper forms
- case workers for complex situations
This approach does come with trade-offs. It increases the cost per interaction, requires more staff time, and can introduce some variability and rework. However, it leads to better overall outcomes: more people receive the benefits they are entitled to, fewer vulnerable users are left out, and the services stay better aligned with its intended policy goals.
If you'd like to explore these ideas further, consult Reimagining the Policy to Service Continuum (TRN5-A02) and Simplifying Government Forms (DDN2-A41). These articles offer practical guidance on improving service performance from a design and a delivery perspective.
Guiding continuous improvement
End-to-end public services are defined by the full outcome. They aim to deliver not just a single step, but a process. For example, it's about getting a passport, not just applying for one. Learning to drive, not just passing a test. Receiving financial support, not just submitting a claim.
Public services are made up of many interconnected parts, such as technology, data, staff tools, service capabilities, and the channels people use to interact with them. Because these parts are interconnected, improving a service is not about fixing one element in isolation. A change in one area often affects other areas. Continuous improvement means looking across the whole service, understanding how these parts work together, and making changes that improve the overall outcome, not just one piece of it.
Services shouldn't be treated like products on a factory line, built once and considered done. Instead, the focus should shift to how well the service is performing over time and how it can be improved. This means asking:
- What level of investment is appropriate to achieve the desired outcomes?
- What trade-offs are we making?
- And where are the risks or missed opportunities?
Looking at service performance from end to end helps teams, projects, and programs understand where they fit and what they should focus on improving. It also provides real evidence to support decisions about what to change and why.
How to start measuring end-to-end service performance
Start by identifying what you want to understand about effectiveness and efficiency. AI tools can help generate ideas, but their outputs should always be reviewed critically.
The most valuable insights are often the hardest to measure, either because organizations aren't used to tracking them, or because the data doesn't exist yet.
Even if you don't have perfect data, improving your understanding, even slightly, puts you in a stronger position to act.
Practical ways to get started
Follow the user
Understand what actually happens when someone uses the service. Track a small number of users or cases from start to finish or ask recent users to describe their experience. This helps identify where things work, where they break down, and why.
Example: Follow 10 recent applicants from submission to receiving a decision to see where delays or errors occur.
Sampling
Regularly review a small set of cases to spot patterns. For example:
- How often was the outcome achieved correctly?
- How confident were users that the right thing happened?
- Where did users get stuck or make errors?
- Did users complete what they set out to do?
Example: Review 50 recent cases each month to identify common errors or incomplete applications.
Use existing contact data
Contact centres and support channels are great sources of data. Look beyond common issues. Track whether problems decrease over time as improvements are made.
Example: Monitor whether calls about a confusing form decrease after it is simplified
Use proxies
If you can't measure the final outcome yet, track something closely related. For example, measure uptake, usage patterns, or task completion instead of long-term impact.
Example: Use application completion rates as a proxy for whether users can successfully navigate the service.
Desk research
Estimate service quality using simple scoring, like usability, usefulness, and usage to help build a case for improvement.
Example: Rate different parts of the service on a scale of 1-10 to identify where investment is most needed.
Analytics and instrumentation
Work with data specialists to track key indicators. Ensure digital tools are built with measurement in mind. Where that's not possible, rely on smaller samples and ongoing testing.
Example: Add tracking to see how long users spend on each step and where they drop off.
Get organized to keep improving
To improve continuously, service performance needs to guide how teams work, make decisions, and prioritize effort.
Build feedback loops
Move beyond one-off learning. Set up ongoing ways to learn and improve. Strong feedback loops act as an early warning system and provide a steady source of evidence for decision-making.
For example:
- Use incoming support requests to identify and fix root causes, not just respond to issues.
- Involve front line staff in improving the tools they use.
- Regularly check whether users achieve the right outcome and feel confident they did.
- Test ideas frequently using simple prototypes.
Team objectives
Use end-to-end service performance to shape team goals and responsibilities.
High-performing teams test, learn, and iterate. They can show what they've changed based on user feedback and evidence. Using common service standards and proven practices can also help teams focus on what works.
Better governance
Use service performance to guide investment decisions and prioritize work across the organization.
New initiatives should clearly explain how they will improve service outcomes and how progress will be measured. Simple, consistent templates can help compare and align efforts across teams.
Role of leadership
Leadership helps connect the full picture of service performance, bringing together policy goals, operational realities, digital delivery and user experience into one shared view. Their role is to keep teams focused on what matters most: improving outcomes for users. This means setting clear priorities, making trade-offs, and challenging assumptions about what will (or won't) work.
Leaders should also create space for learning by encouraging teams to test ideas, use evidence, and adjust course when needed. AI tools can support this by helping explore options or questions and assumptions, but should always be used with critical judgment.
Finally, leaders can drive alignment by asking for simple, end-to-end service dashboards, even early prototypes. These don't need perfect data, but they help highlight what is known, what is missing, and where to focus improvement efforts.
A starting plan
No service will ever be perfectly set up to measure everything you want.
Instead, focus on making steady progress towards a clearer and more consistent approach to performance while building capability to act on what you learn.
Start with a few practical steps:
- Agree on what matters
Don't focus on what's easy to measure. Be clear on what's important, even if you can't track it yet.
- Map your services
Define your end-to-end service. Who's involved, what's included, and where work overlaps. This helps create a service view shaped by performance.
- Design with measurement in mind
Make instrumentation and testing standard practice. Encourage teams to prototype, test ideas, and assess risks. Ask what they want to measure, and how it connects to overall service performance.
- Plan in sequence
Be clear on priorities: what can you do now, what comes next, and what will take longer. For example, a simple dashboard might be quick to deliver, while improving supplier data may take more time.
Conclusion
Measuring public service performance from end to end isn't about establishing perfect metrics or adding excessive new layers of reporting. It's about developing a clearer, shared understanding of what "good" looks like in context and using that to make better-informed decisions over time.
No service starts in a perfect position. But we can continuously improve by reflecting on the effectiveness and efficiency of our services and users' experiences interacting with it. We can employ approaches such as sampling, research, and using proxies to better assess whether services are truly delivering their intended outcomes.
Resources
