Skip to content
Design Academy My progress

Stage 2 · Lesson 13 of 13

Defining success metrics

Choosing measures that show whether a design actually helped.

  • Beginner
  • 30 min
  • 3 check questions
  • Exercise: 35 min

By the end of this lesson you can

  • Define what success looks like for a design change before you build it.
  • Choose metrics that reflect user outcomes, not just activity.
  • Combine behavioural and attitudinal measures, and watch for unintended effects.

The idea

If you do not decide what success looks like before you start, you will always find a number afterwards that looks good. Success metrics make the goal concrete and honest.

Outcomes, not activity

Page views, clicks and time on site are easy to measure but often misleading. Longer time on a help page might mean people are stuck. Good metrics reflect whether people achieved what they came to do: task completion rate, time to complete a task, error rate, successful first-time setup, repeat use.

Frameworks

Google's HEART framework groups metrics into Happiness, Engagement, Adoption, Retention and Task success. For each category you choose a goal, a signal that would show progress, and a metric you can track. You rarely need all five; pick the ones that fit the problem.

Behaviour and attitude

Behavioural metrics measure what people do (completion, errors). Attitudinal metrics measure what they say (satisfaction scores, perceived ease). Use both: people sometimes complete a task but hate the experience, or enjoy it but fail.

Guardrail metrics

A guardrail metric checks that you have not broken something else. If you reduce checkout steps to increase conversion, watch refund requests and support calls too.

Why it matters in real projects

Metrics turn design from opinion into something the team can learn from. They help you prove the value of your work, decide whether to keep a change, and spot harm early. Defining them up front also keeps everyone honest: success is agreed before results come in.

A realistic example

A water company redesigns the "report a leak" form. The team defines success before launch:

  • Task success: share of started reports that are submitted rises from 61% to 80%.
  • Quality: reports with a usable location rise from 70% to 90% (fewer callbacks to engineers).
  • Happiness: average ease rating after submitting is at least 4 out of 5.
  • Guardrail: duplicate reports do not increase.

After a month, completion reaches 83% and location quality 88%, but duplicates rise slightly. The team adds a "Has this already been reported?" map, and duplicates fall again. Without the guardrail they would have missed it.

Step by step

  1. Restate the goal

    Write the user and business outcome you want in one sentence.

  2. Choose categories

    Use HEART or similar to pick the two or three that matter most.

  3. Define goal, signal, metric

    For each category, write what you want, what would show it, and how you will measure it.

  4. Record a baseline

    Measure the current value before you change anything.

  5. Set a target and timeframe

    Decide what change counts as success and over what period.

  6. Add a guardrail

    Pick one metric that would reveal unintended harm.

Common mistakes

  • Mistake Choosing vanity metrics such as page views.

    Instead Measure whether users achieved their goal.

  • Mistake No baseline.

    Instead Measure before the change so you can compare.

  • Mistake Defining success after launch.

    Instead Agree metrics and targets before you build.

  • Mistake Ignoring side effects.

    Instead Track at least one guardrail metric.

Check your understanding

Choose an answer, then check it. There is an explanation either way.

Question 1 of 3Time on a help page increases after a redesign. What could this mean?
Show answer

It could mean people are more engaged or more stuck; you need other evidence. Activity metrics are ambiguous without outcome measures.

Question 2 of 3What does the "T" in HEART stand for?
Show answer

Task success. Task success measures whether people achieve what they set out to do.

Question 3 of 3What is a guardrail metric for?
Show answer

Detecting harm caused elsewhere by a change. Guardrails catch unintended side effects.

Practice exercise

Exercise · about 35 min

Metrics plan

Define success metrics for a redesign of a cinema ticket booking flow.

Steps

  1. Write the goal in one sentence.
  2. Choose two or three HEART categories and fill in goal, signal and metric.
  3. Note how you would get a baseline.
  4. Add one guardrail metric and explain what it protects.

Deliverable

A one-page metrics plan with targets and a guardrail.

Self-assessment

Tick each item when it is true of your exercise. Be honest; this is for you.

4 items

Summary

  • Define success before building, with baselines and targets.
  • Measure outcomes such as task success, not just activity.
  • HEART helps structure goals, signals and metrics.
  • Guardrail metrics reveal unintended harm.

Further reading

Third-party sources, chosen for reliability. They open in a new tab.

Your notes

Marking complete updates your course and roadmap progress.

Back to UX research and problem definition Roadmap

PreviousHypotheses, assumptions and prioritisation Next lessonOrganising information