dev-skills

Appearance

HEART Metrics ​

What this topic is for ​

Measuring user experience at scale is hard. Traditional usability methods — lab studies, task completion rates, user satisfaction surveys — work well for small samples but don't scale to millions of users. Logging behavioral data at scale works, but without a framework it produces hundreds of possible metrics with no guidance about which ones matter or what they mean together.

HEART is a framework for selecting user-centered metrics for large-scale web products. It organizes the space of possible metrics into five dimensions — Happiness, Engagement, Adoption, Retention, and Task Success — each covering a different aspect of the user experience. Choosing metrics from across these dimensions lets a product team measure the health of their product without drowning in undifferentiated data.

This topic covers what each HEART dimension measures, how to measure it, and how to decide which dimensions matter for a given product. For guidance on translating dimensions into specific metrics, see Goals–Signals–Metrics Process.

The five dimensions ​

Happiness ​

Happiness measures user attitudes: how users feel about the product. This includes satisfaction, perceived ease of use, perceived quality, and willingness to recommend. It captures what users think and feel, not what they do.

Happiness is almost always measured through surveys — satisfaction questions, ease-of-use ratings, or the Net Promoter Score (NPS). Because survey data is sparse and delayed compared to behavioral logs, happiness metrics update less frequently than other HEART dimensions, but they capture signal that behavior alone cannot: a product can show high engagement while users hate it (think: social media compulsion loops). Happiness keeps that from being invisible.

Common happiness metrics:

  • User satisfaction score (e.g., a 5-point "how satisfied are you?" survey item)
  • Perceived ease of use (e.g., "how easy was it to complete your task?")
  • Net Promoter Score (NPS)
  • Sentiment from open-ended survey responses

For happiness data to be useful, the survey must reach a representative sample. In-product surveys tied to specific interactions — for example, shown after a user completes a task — are more reliable than periodic blanket surveys, which tend to over-represent power users.

Engagement ​

Engagement measures the depth of user involvement with a product. High engagement means users are interacting frequently, spending more time, or using a wider range of features. What counts as engagement depends on the product: for a social network, engagement might be daily active use and number of posts; for a document editor, it might be edit sessions per user per week.

Engagement is measured from behavioral logs. Common metrics:

  • Visit frequency (e.g., visits per user per week)
  • Session length (e.g., average time per session)
  • Feature breadth (e.g., number of distinct features used in the past 30 days)
  • Depth per session (e.g., photos uploaded per active day)

The right engagement metric is product-specific. A tax tool used once a year shouldn't be measured by visit frequency; a messaging app should. Choose engagement metrics that reflect the intended usage pattern for your product.

Adoption ​

Adoption measures new user growth: how many users are taking up the product or a new feature in a given time period. This applies to the product overall (new accounts) or to features within a product (users who've tried a feature for the first time).

Common adoption metrics:

  • Number of new accounts created per month
  • Number of users who have tried a feature at least once
  • Upgrade rate to a new version
  • First-use rate for a new feature among existing users

Adoption is about breadth and reach. A product with high adoption attracts many users; a feature with high adoption gets discovered and tried. Adoption alone doesn't say whether those users stay — that's what Retention covers.

Retention ​

Retention measures whether users come back. A high-retention product keeps its existing users over time; a low-retention product loses users as fast as it gains them.

Common retention metrics:

  • N-day retention: what fraction of users who first used the product on day 0 return on day N? (7-day and 30-day are common)
  • Churn rate: fraction of users who stop using the product in a given period
  • DAU/MAU ratio: daily active users divided by monthly active users — a measure of how "sticky" a product is
  • Feature retention: do users who try a feature come back and use it again?

Adoption and Retention work together as a pair: Adoption tells you how many users are starting; Retention tells you how many are staying. A product with high Adoption and low Retention is a leaky bucket; a product with high Retention but low Adoption is healthy for its existing users but not growing.

Task success ​

Task success measures traditional UX metrics at scale: whether users can accomplish what they came to do. It covers efficiency (how long a task takes), effectiveness (whether it completes), and error rate. Task success differs from Engagement in direction: high Engagement means users are doing more; high Task Success means users are doing it better — faster, with fewer failures, with less effort.

Common task success metrics:

  • Task completion rate: did users finish what they started?
  • Time to completion: how long did the task take?
  • Error rate: what fraction of attempts encountered an error?
  • Search success rate: what fraction of searches led to users finding what they needed?
  • Abandonment rate: what fraction of users gave up mid-task?

For products with clear task flows — search, checkout, form submission — Task Success is usually the most important dimension to measure. For ambient or social products without defined task flows, it matters less.

Choosing which dimensions to focus on ​

Not every product needs metrics from every HEART dimension. Teams that measure everything end up with a dashboard no one reads. The goal is to select a small, coherent set of metrics that together tell the story of the product's health.

Start with the product's purpose. A search engine's core job is to help users find what they're looking for — Task Success is central. A social network's core job is to keep people connected — Engagement and Retention dominate. A new product trying to reach more users might weight Adoption heavily. Let the product's reason for existing drive which dimensions matter most.

Ask what problem the team is trying to solve. If users are completing tasks but saying they hate the product, Happiness deserves attention. If growth is stalling, Adoption and Retention are diagnostic. Metrics should connect to what the team is actually trying to change.

Don't ignore dimensions just because they're hard to measure. Happiness is measured through surveys, which require operational overhead. That overhead is often worth it — survey data catches problems that logs cannot.

Keep the set small. Two to three dimensions, with two to three metrics each, is usually enough. A team watching 20 metrics is watching none.

Further reading ​