
CX Measurement
Customer experience for digital services
Assess digital service CX by following a customer task from entry to confirmed outcome, removing avoidable steps and checking the service behind the screen.
A good digital service helps a customer finish the task they came to do, understand the outcome and get help when the online path fails. Assess the experience from the customer’s starting point to a confirmed result. A completed form is not necessarily a completed service.
Start with a task, not a screen
Choose a specific need, such as changing contact details or submitting a request. Define what success means to the customer: the change takes effect, the request is received, or the next step is clear. Trace the path from finding the service to receiving confirmation. Include messages, waiting periods and work the customer must do outside the website.
Observe people attempting the task on devices and in circumstances that reflect the intended audience. Include people with different access needs and levels of digital confidence. Give them a goal without telling them where to click. Record where they hesitate, backtrack or ask for help before deciding why.
A session can expose a failure worth fixing. It cannot establish how common that failure is across all customers.
Give people enough context before they enter a transaction to recognise whether it may apply to them. Broad eligibility information, such as an age or location requirement, can sit on a start or guidance page.
Handle more complex eligibility through questions in the transaction, using answers to determine whether the person can continue and what they should do next. This helps people get advice relevant to their circumstances without having to interpret complicated criteria themselves, and can reduce time and money spent processing queries.
Remove avoidable decisions
At each step, ask whether the customer needs to make a decision or supply information to achieve the task. A field may be essential to fulfil a request, while an account prompt or internal category may serve mainly the organisation. Explain necessary checks when they occur, including what happens after submission.
An account can help someone return regularly to view or update information. For a one-off task, consider whether the customer can proceed without one and still check the outcome securely. If an account is necessary, ask for it when its purpose becomes clear. The choice depends on the service’s access, security and follow-up needs.
For a one-off transaction, a unique reference can let someone check its status with their name or email address instead of creating an account. Because references can be hard to remember, send them by email or text message. If an account is needed, make clear the difference between creating one and signing in, and avoid asking for the same information twice.
Put useful help where the question arises
Help should answer the question blocking the next action. A short explanation beside an unfamiliar field may do more than a long help article.
Tell customers what information is needed, why it is needed where relevant, and how to correct an error. If people still cannot proceed after reading the explanation, examine the question or process itself. More text cannot repair an impossible requirement.
Test help in the same task and at the moment of confusion. Watch whether people notice it, understand it and complete the step correctly. Check that instructions and error feedback are available to people using a keyboard or assistive technology.
Connect the screen to the service behind it
A clear interface cannot compensate for a request that reaches the wrong team or produces no usable confirmation. When research shows a problem, follow the request through the systems and people responsible for the outcome. Decide whether the change belongs in wording, interface behaviour, a service rule or the work behind the screen. Give each change an owner and a way to check its effect.
For a live service, review task completion alongside errors, support contacts and customer feedback. Keep the measures tied to the same task and consider who may be missing from the data, including people who leave before submitting. Revisit the journey after a change: customers may pass the edited screen yet meet the same obstacle later.
Check that decisions made in the transaction lead to the right service action and that people can see what happens next. This connects the on-screen interaction to the operational work that follows.
The Mandarin reported that Adobe’s Digital Government Index for Australia fell to 67.7 out of 100 in 2024, a one per cent decrease on the previous year. Customer Experience improved modestly, while Site Performance and Digital Equity declined.
Australia's Digital Government Performance (2024)
- Digital Government Index Score
- 67.7 out of 100
- Customer Experience Trend
- Modest improvement
- Site Performance Trend
- Declined
- Digital Equity Trend
- Declined
In this guide
- Observing where users struggle with an online taskPlan a neutral online task, record moments of struggle and turn observations into bounded usability findings.
- Comparing account creation with guest access from a CX perspectiveCompare account and guest routes by task completion, return access, recovery and service security needs.
- Testing help content at the point of confusionFind where customers need help, test an explanation in the task and check whether it improves correct completion.
- Connecting usability findings with service redesignTrace a usability finding to its service cause, assign an owner and check whether the redesign improves the customer outcome.



