What needs to be finished first
The product has tracking infrastructure (Mixpanel, Amplitude, Heap, etc.) with 60-90 days of event data.
What needs to be ready before start
Whiteboard or Miro board with five sections (Happiness, Engagement, Adoption, Retention, Task Success); GSM table (Goals, Signals, Metrics) as template; list of existing metrics; dashboard access; happiness survey tool (Wootric, Delighted, or in-house survey).
A UX lead or product manager as owner; data analyst for metric availability; designer for UX perspective; engineer for tracking feasibility; optional researcher for happiness measurement.
Product strategy and goals; existing metrics; available tracking events; user segments; channels for survey distribution; comparable HEART implementations as references.
120 min initial, ongoing after that
Create five sections on the board with columns Goals, Signals, Metrics. Limit one to two goals per dimension. Limit one to two signals per goal and one metric per signal. More creates cognitive overload.
The one question this method answers
Which five UX dimensions do we measure with which goals, signals, and metrics, and how do we connect them into balanced decision-making?
Marker: Phase
| Step | Duration | Action | Hint |
|---|---|---|---|
1Phase 1: Define one goal per dimension | 30 min | Define exactly one goal per HEART dimension, for example Happiness = satisfaction goal, Engagement = usage intensity goal, Adoption = uptake goal, Retention = return goal, Task Success = success goal per key task. | A goal is a qualitative statement, not a metric. "Users recommend the tool to colleagues" is a goal, "NPS > 40" is a metric. |
2Phase 2: Identify signals | 30 min | For each goal define one to two observable signals: what changes in behavior or statements when the goal is reached? For example, engagement signal: sessions per week. | Signals are not metrics. A signal is an observable behavior change; the metric is the number. One metric per signal, no more. |
3Phase 3: Specify metrics | 30 min | For each signal define metric with numerator, denominator, period, source, and baseline. For Happiness, define survey question and distribution channel. | Keep to one to two metrics per dimension. Defining five metrics per dimension turns HEART into a vanity dashboard. Keep metrics minimal. |
4Phase 4: Tracking setup and review cadence | 30 min | For each metric check whether it already exists (dashboard link), must be built (engineering ticket), or is survey-based (tool setup). Define review cadence per metric. | Survey metrics need distribution cadence (for example weekly to 5% of active users). Behavioral metrics run continuously. Review weekly for operational tracking or monthly for strategy. |
What comes out at the end
GSM table per dimension with goals, signals, metrics, baseline, target, data source, review cadence. Dashboard with all five to ten metrics as a HEART heatmap. Happiness survey setup.
Re-evaluate GSM table quarterly. Version metric definitions (for example on tracking changes). Keep survey questions stable or lose comparability. Archive dashboard snapshots monthly.
Compact working template for HEART Framework with context, input, output artifacts, and next step.
What is this method used for?
Which question should be answered at the end?
Which data, observations, or materials are available?
Owner, date, success signal.
Concrete filled scenario, fictional example
Concrete filled scenario, fictional example
Happiness
Engagement
Adoption
Retention
Task Success
Review cadence: Weekly adoption and task success in team standup, monthly review across all five metrics.
Recognize symptoms and steer against them
Five to seven metrics per dimension, and no one can track them all.
Rule of thumb: one to two metrics per dimension. If more are needed, create a sub-dashboard. HEART is for overview, not exhaustiveness.
Happiness metric is derived from behavioral data (for example active users = happy users), so actual mood is not measured.
Happiness requires direct questioning. Set up survey question and distribution cadence. Use NPS or CSAT as standard, with custom questions for niche contexts.
GSM table has goal and metric in one cell, while signal column is missing or repeats metric.
Signal is observable behavior change; metric is the number. Name signal first, then metric for each goal. Keep this separation.
Task Success metric is "app works" with no clear connection to a specific task.
Define one to three core tasks per product. Measure task success per main task separately. Generic "success" is not measurable.
Metrics are measured but not linked to backlog levers.
Define top levers per dimension each quarter. Tag backlog items with HEART dimensions. HEART should drive roadmap priorities, not just reporting.
Done signals checkable in under a minute
Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.