By Mehdi HassanWhich UX metrics should a SaaS product track?
A SaaS product should track UX metrics that show whether people get value, complete their key tasks and come back. For most products, that means activation rate, task success rate, time to complete the core task, feature adoption, retention and a simple satisfaction measure. Two or three well-defined metrics beat twenty vague ones.
UX metrics differ from business metrics such as revenue or churn. They measure the experience that drives those outcomes, so they move earlier and tell you more about what to fix. A drop in task success on the invite flow usually shows up well before a drop in seats per account.
This guide is part of our user research for startups playbook. Its angle is practical: every metric below comes with the analytics event your developers need to instrument, so the metric exists in your data rather than only in a slide.
What is the HEART framework for UX metrics?
The HEART framework is a model developed at Google for measuring user experience at scale. It groups UX metrics into five categories: Happiness, Engagement, Adoption, Retention and Task success. Teams pair it with a process called Goals, Signals, Metrics, which turns each category into a specific, measurable number.
Category | What it measures | Typical SaaS metric |
|---|---|---|
Happiness | How users feel about the product | Satisfaction score from an in-app survey |
Engagement | How much and how often people use it | Active days per week for key users |
Adoption | How many people start using a product or feature | Share of new accounts that use a feature in their first weeks |
Retention | How many people keep coming back | Share of accounts active again after a set period |
Task success | Whether people complete key tasks efficiently | Completion rate and time for the core flow |
You don't need all five. Pick the categories that match your current goal. An early-stage product focused on onboarding usually cares most about adoption and task success. A mature product fighting churn cares about retention and happiness.
How do you turn a UX goal into a metric?
You turn a UX goal into a metric with the Goals, Signals, Metrics process: write the goal in plain words, name the user behavior that would signal progress, then define the exact number you'll track. Skipping straight to metrics usually produces numbers that are easy to collect and meaningless to act on.
Goal
Write what you want users to achieve, in a sentence. For example: new admins should get their team working in the product quickly.
Signal
Name the behavior that shows the goal is being met. For example: an admin invites teammates and at least one of them accepts.
Metric
Define the number precisely. For example: share of new workspaces with at least one accepted invite within seven days of sign-up.
Event
Name the analytics events that produce the metric: workspace_created, invite_sent and invite_accepted, each with a workspace id and timestamp.
Target and review date
Set a direction or a target and a date to look at it, so the metric actually informs a decision.
Which analytics events does each UX metric need?
Each UX metric needs a small set of named events with consistent properties, defined before the feature ships. Write them in a shared tracking plan so designers, developers and anyone reading the dashboard agree on what each number means. Most product analytics tools work with this event-and-property model.
UX metric | Events to instrument | Key properties |
|---|---|---|
Activation rate | signed_up, activation_event (your value moment) | account id, plan, time since sign-up |
Task success rate | task_started, task_completed, task_failed | task name, step reached, error type |
Time on task | task_started, task_completed | timestamps, task name |
Feature adoption | feature_used (per feature) | feature name, account id, first use flag |
Retention | session_started or key action events | account id, user id, date |
Error rate | error_shown | error code, screen, recoverable or not |
Satisfaction | survey_shown, survey_answered | score, comment, screen where asked |
Keep event names consistent: object then action, in the past tense, in one case style. 'invoice_sent' and 'invoice_paid' are easy to query. 'SendInvoiceClick' and 'paid-invoice' in the same plan are not.
For example, suppose you instrument task success for an invoice flow
For example, suppose the core task in your product is sending an invoice. You define it as starting when the user opens the new invoice form and ending when the invoice is sent.
Events: invoice_form_opened (with source: dashboard, client page or command menu), invoice_validation_failed (with field name), invoice_sent (with time since form opened) and invoice_form_abandoned (fired when the user leaves the form without sending, with the last field touched).
Task success rate is invoices sent divided by forms opened. Time on task comes from the time-since-opened property. The validation event tells the designer which fields cause trouble, and the abandoned event tells them where people give up. One well-specified flow gives you three or four UX metrics from a handful of events.
How do you measure task success and time on task?
Measure task success as the share of attempts that reach a defined end state, and time on task as the time between a defined start and that end. Both depend on agreeing exactly where a task starts and finishes. Without that, two people will calculate different numbers from the same data.
- Define the start. The first action that commits to the task, such as opening the form, not merely landing on the page.
- Define success. The state that means the job is done, such as the invoice being sent, not just saved as a draft.
- Define failure. An error that blocks completion, or abandonment after a set period of inactivity.
- Report the median time. A few people who leave a tab open overnight will distort an average.
- Segment. New users and experienced users often have very different numbers, so look at them separately.
Usability testing gives you the same metrics in miniature: whether each participant completed each task and how long it took. Our guide to usability testing on a budget shows how to record them in a session and score the issues behind them.
How do UX metrics connect to activation and onboarding?
UX metrics connect to activation by measuring how many new users reach the moment they first get value, and how quickly. Activation is often the single most important UX metric for an early-stage SaaS product, because everything after it, from retention to revenue, depends on people getting there.
Define your activation event carefully. It should be a behavior that predicts people sticking around, such as sending a first invoice or sharing a first report, not simply finishing a setup checklist. Our guide to SaaS onboarding UX covers how to design the path to that event, and our guide to UX writing for SaaS shows how copy in empty states and prompts moves it.
If your marketing site feeds sign-ups, measure the handover too. Our guide to SaaS conversion rate benchmarks covers visitor-to-trial and demo rates, and our guide to website redesign goals and KPIs explains how to set targets for a site relaunch.
How many users do you need before UX metrics are useful?
UX metrics become reliable once enough people go through a flow that a change in the number isn't just noise. With a few dozen users, a swing of several percentage points can come from a single account. That doesn't make metrics useless early on, it means you read them as hints and confirm them with qualitative research.
Pros
- Events set up early give you a baseline before big changes
- Even small numbers can reveal a step where nobody gets through
- Instrumenting forces the team to define tasks and success clearly
- Metrics scale with you, so the work pays off as users grow
Cons
- Percentages on small samples swing a lot from week to week
- Metrics show what happened, not why
- Too many events create noise and slow analysis
- Poorly named events need rework later
The practical answer is to instrument a small set of events now and pair them with regular usability tests until your numbers are large enough to trust on their own. For AI features, where success is harder to define, our UX readiness checklist for adding AI to a SaaS product covers what to measure. Our brand & design sprint, at $3,000 fixed, is a good next step once your metrics show which flow needs redesigning.
