Mixpanel Product Analytics Certification flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

Event

A tracked user action with a name, a user identifier, a timestamp, and optional properties. All funnels, retention, and segmentation analyses are queries over the event stream.

Distinct ID

The canonical identifier for a user in the project. Events sent with the same distinct ID are attributed to one user across sessions and devices.

Identify Call

A client call that links the current anonymous ID to a known user ID. It merges pre-login activity into the identified profile so earlier events are not orphaned.

Funnel Conversion Rate

Users who completed the final step divided by users who completed the first step, within the configured conversion window. Step order and window length change the result.

Retention Cohort

A group of users sharing the same first-activity date. Retention is measured as the share of that cohort active again on day N after their start.

DAU/MAU Stickiness

Daily active users divided by monthly active users. A high ratio means users return frequently; a low ratio means usage is occasional or seasonal.

Segmentation Breakdown

Splitting a metric by an event property or user profile property. The property must be sent with the event or stored on the profile to appear as an option.

North Star Metric

A single leading indicator of delivered customer value that correlates with long-term retention, such as weekly active teams completing a core action.

Pageview Event

An event fired when a page loads. Single-page applications must fire it explicitly on route change because no full document reload occurs.

UTM Parameters

Five query parameters (source, medium, campaign, content, term) appended to a URL. They are captured as properties and drive campaign attribution in reports.

First-Touch Attribution

Credit for a conversion assigned to the earliest campaign that acquired the user. It rewards acquisition channels and understates channels that assist later.

Last-Touch Attribution

Credit for a conversion assigned to the most recent campaign before the conversion. It rewards closing channels and can overstate retargeting performance.

Direct Traffic

Sessions with no referrer and no UTM parameters. Bookmarks, untagged email links, and privacy-stripped referrals all land in this default bucket.

Session Timeout

The inactivity window, commonly 30 minutes, after which a new session starts. Session-scoped metrics such as sessions per user depend on this setting.

Bounce Rate

Sessions containing only a single tracked event divided by total sessions. It rises artificially when a landing page fires just one event type.

Cross-Domain Tracking

Configuring the same project token and identity passing across domains so one user is not split into separate distinct IDs with fragmented histories.

Mobile Event Queue

The SDK's local buffer that batches events before sending. Events not flushed before the app is killed can be lost permanently.

App Lifecycle Events

Install, update, open, and background events auto-collected by the mobile SDK. They form the base for activation and retention cohorts on mobile.

Push Open Rate

Opened push notifications divided by delivered notifications, not by sent. Delivery failures and bounces must be removed from the denominator.

Deferred Deep Link

A link that survives an app install and resolves on first launch, letting a new install be attributed to the campaign that drove it.

App Version Breakdown

Segmenting a metric by app version to compare behavior before and after a release. It is the fastest way to isolate a release regression.

Advertising ID Opt-In

User consent governing access to IDFA or GAID. When opt-in is low, device-level mobile attribution becomes probabilistic and needs a match-rate caveat.

Install Attribution

Assigning a new app install to the ad click or campaign that caused it. On mobile the app store sits between the click and the first launch.

Randomization Unit

The entity randomly assigned to control or variant, such as user, device, or session. It must match the unit used in the analysis or results are biased.

P-Value

The probability of observing a difference at least this extreme if the null hypothesis is true. It is not the probability that the variant is better.

Statistical Power

The probability of detecting a real effect of a given size, conventionally set at 80%. Underpowered tests miss true effects and overstate winners stopped early.

Minimum Detectable Effect

The smallest lift a test is designed to detect. Required sample size grows roughly with the inverse square of this value.

Peeking Problem

Checking results repeatedly and stopping at the first significant reading. It inflates the false-positive rate above the stated threshold.

Guardrail Metric

A secondary metric such as crash rate, latency, or revenue per user that must not degrade. It catches hidden costs of a primary-metric win.

Novelty Effect

A temporary behavioral shift caused by a feature being new rather than better. It distorts early experiment readings and fades over subsequent weeks.

Session Replay

A reconstruction of a user session from DOM snapshots and input events. It is a compressed event log replayed in a browser, not a video recording.

Input Masking

Rules applied at capture time that redact text, form inputs, or blocked elements. Data not masked before transmission cannot be recovered for privacy compliance.