Good experience measured by what
Experience can be measured, but measuring incorrectly is worse than not measuring at all.
Good experience is measured by two groups of metrics: behavioral (did the customer complete the task, how long did it take, how many errors, did they return) and perceptual (did the customer find it easy or difficult, how satisfied were they). You don’t need to measure everything, just select the group linked to the page's goals: for a page with a clear call to action, look at completion and conversion rates; for repeat-use products, add retention and ease of use. The most important thing is to have a baseline before making changes; otherwise, the numbers after completion will mean nothing.
- page with clear action goals such as buy, register, schedule
- Digital products that are repetitive need to measure retention and satisfaction.
- traffic is still low, need to use surveys for perception instead of behavioral data
- want to measure all metrics at once, with no priorities
- there is no baseline before making changes for comparison
Good experience is not an ambiguous feeling. It can be measured, and when measured correctly, you will know if the design is effective or just looks good. The issue is not whether to measure or not, but choosing the wrong things to measure and getting overwhelmed by numbers that do not lead to any decisions.
Two groups of key metrics to know
There are two groups of metrics to evaluate the experience, and both are necessary as they complement each other, not replace each other.
- Behavior metrics: Whether customers complete tasks, how long it takes, how many errors occur, the conversion rate (the percentage of people taking the target action such as purchasing, signing up, or scheduling), and whether they return. These numbers come from real data, not self-reported by customers.
- Perception metrics: Customers self-report how easy or difficult they find things, and their level of satisfaction. Common tools include the SUS (System Usability Scale, a standardized 10-question usability measure) and SEQ (Single Ease Question, a single ease question asked immediately after a customer completes a task).
The reason both are needed: customers may express satisfaction, but their behavior shows they are taking twice as long as necessary to find the payment button. Looking at just one group is only seeing half the picture.
Choose metrics based on goals, not measuring everything.
The HEART framework proposed by Kerry Rodden and colleagues at Google (published at the ACM CHI 2010 conference) suggests five user-centered metrics: Happiness, Engagement, Adoption, Retention, and Task success. The key point of this framework is not to use all five metrics, but to choose which ones align with the current project goals.
Recurring digital products (applications, management platforms, internal tools): focus on retention and ease of use. For this type of product, customer return is the true measure, not first-time use.
Baseline: an essential element.
The number after completion only has meaning when compared to a previous number. If you only measure once after launching a new version, you won't know if you've improved or regressed. The correct process is to measure before making changes, record that baseline, and then measure again after implementation.
When traffic is still low and there is not enough behavioral data to trust, conducting perception surveys like SUS or SEQ is a reasonable way to start. MeasuringU compiles data from various studies and shows that the average SUS score of samples is around 68, but this average depends on the context and is not an absolute threshold for you to set a rigid target.
Common errors when measuring experiences
- Measure all metrics at once and then feel overwhelmed; no metric truly leads to design decisions.
- Only look at page views or time spent while overlooking the completion rate. Many views do not mean many people are accomplishing what they need to do.
- Fully trust customer satisfaction scores provided without cross-referencing with actual behavior. Customers often report better than their actual experiences.
- Only measure once after completion, without a baseline for comparison.
The viewpoint of Sinh Vũ
Linking experience metrics to business outcomes allows leadership to see the value. Don’t report on design activities; report on what has changed in revenue, support costs, or retention rates.
Nielsen Norman Group, Stop Reporting UX Activity and Report Business Outcomes
Sinh Vũ believes that experience must be measurable, but choose a few metrics tied to your goals instead of covering the entire scoreboard. For pages with clear calls to action, Sinh Vũ tracks completion rates and conversions. For repeat-use products, add retention and ease of use. Sinh Vũ does not promise a specific conversion number before starting because that number depends on many factors beyond design, but always establishes measurement methods and baseline milestones so you know if your investment yields results.
Decision checklist
Topic: How to measure a good experience. Sinh Vũ guide, sinhvu.com
Select each item you find appropriate, then print or save as PDF to take with you.
If you have marked most of the signs above, this is the time to discuss in more detail. Sinh Vũ can help you review and propose a direction.
References
Kerry Rodden, Hilary Hutchinson, Xin Fu. Measuring the User Experience on a Large Scale, ACM CHI 2010 (HEART framework). Nielsen Norman Group, Stop Reporting UX Activity and Report Business Outcomes. MeasuringU (Jeff Sauro), 10 Essential Usability Metrics and a summary of SUS benchmarks.
Frequently asked questions
A SUS score of 68 means my page meets the standard, right?
A score of 68 on the SUS (System Usability Scale) compiled by MeasuringU from various studies is often mentioned as the average of the sample, not an absolute threshold. You should use this number to compare before and after making changes, or against your own previous version, rather than viewing it as a pass or fail. The industry context and user demographics greatly influence this number.
I just launched and have low traffic. Is measuring behavioral metrics reliable?
When traffic is still low, behavioral data can be noisy due to the small sample size. During this phase, Sinh Vũ advises you to use perception surveys like SUS or ease-of-use questions after each task (SEQ) to compensate. As traffic increases, these two data sets should be compared together because customers may express satisfaction, but their behavior may indicate otherwise.