Pendo

Pendo exposes several object types through its aggregation API: visitors, accounts, pages, features, track events, guides, and NPS poll responses. This integration focuses on three of these — feature events, guide events, and visitor metadata — because they are the objects most directly shaped by the product team's decisions. Account-level aggregates and page-view events were excluded: they either lack individual attribution or serve as vanity indicators disconnected from the team's contribution. Six KPIs were retained, selected against three criteria: ability to attribute to an identified visitor, resistance to gaming, and balance between leading signals of activation and lagging confirmation of engagement.

Feature adoption: measuring discovery and activation speed

Feature Adoption Rate measures the proportion of visitors who clicked a given feature at least once during the period. Time to First Use measures the average number of days that elapse between a visitor's exposure to an onboarding guide and their first interaction with the feature that guide promoted. These two indicators address the same phenomenon — feature discovery — from two different angles.

Feature Adoption Rate answers the question of whether a feature is being reached by the user population. Time to First Use answers the question of how long it takes a user to act after being exposed to information about the feature. A high adoption rate combined with a long Time to First Use suggests that discovery is happening passively, without the guide playing a causal role in activation — the feature is being found through other means, or users are returning to it after forgetting the guide. A low adoption rate combined with a short Time to First Use identifies the opposite pattern: users who follow a guide act quickly, but most users are never exposed to that guide in the first place, pointing to a targeting or placement problem rather than a content problem.

Feature Adoption Rate carries a gaming risk: a product manager can artificially raise it by surfacing a feature prominently in navigation without demonstrating any genuine value delivery. For this reason, it is read in conjunction with Days Active and NPS Score, which are described below. A rise in Feature Adoption Rate unaccompanied by a sustained engagement signal or satisfaction improvement suggests cosmetic discovery rather than genuine adoption.

Onboarding effectiveness: reading guide outcomes as a pair

Guide Completion Rate measures the proportion of guide views that result in a visitor completing the full guide sequence. Guide Abandonment Rate measures the proportion of guide views that result in the visitor dismissing the guide before completion. These two indicators are structurally paired: they share the same denominator (guideSeen events) and together account for almost all outcomes of a guide exposure.

The reason for tracking both rather than relying solely on completion rate is that dismissal behavior carries information that non-completion does not. A visitor who does not complete a guide may have navigated away, minimized the window, or simply left the browser tab open — none of which signals friction. A visitor who explicitly dismisses the guide signals that the guide was perceived as an interruption. A high Guide Abandonment Rate therefore identifies a specific and actionable problem: guide relevance, placement, or timing is generating resistance, not merely indifference. This distinction is invisible when only the completion metric is available.

Guide Completion Rate is classified as a lagging indicator because it reflects the outcome of an onboarding flow that has already been deployed. Guide Abandonment Rate operates as a leading signal: a rise in dismissal rates on a recently modified guide is detectable before the next feature adoption or retention cycle closes, providing an early correction opportunity.

Engagement depth and satisfaction: lagging validation signals

Days Active measures the average number of days a visitor was active in the product during the reporting period. NPS Score captures the latest satisfaction rating submitted by a visitor through an in-app poll, on a scale of zero to ten. Both are lagging indicators: they confirm patterns that leading indicators have already suggested.

Days Active serves as the primary counter-indicator to Feature Adoption Rate. A product team that introduces a new feature can raise the adoption rate in the short term through in-app guidance and navigation changes. If that feature delivers genuine value, Days Active should hold stable or increase as users return to the product to use it repeatedly. If Days Active remains flat or declines while adoption rate rises, the feature is attracting single interactions rather than building a habit — a pattern that predicts churn regardless of what the adoption metric suggests.

NPS Score provides a different dimension of validation. It captures perceived value rather than behavioral frequency: a user can be highly active in a product they find frustrating, or infrequently active in a product they value highly. The combination of Days Active and NPS Score distinguishes between these two cases. A visitor with rising Days Active and a declining NPS score is a retention risk despite behavioral engagement — a signal that product quality issues are accumulating beneath an apparently healthy usage pattern.

Scope and limits of the integration

Pendo's data model is visitor-centric: every event belongs to a visitorId, not to a product team member. The attribution chain from a KPI to a specific product manager's contribution is therefore indirect. Feature Adoption Rate and Guide metrics are attributable to the PM who authored the relevant feature tag or guide, but only if the team maintains a consistent mapping between Pendo objects and individual owners. Without that operational discipline, the KPIs measure product-level outcomes without identifying who is responsible.

Additionally, the integration depends on the customer's Pendo configuration. If visitorId is set to the user's email address — which Pendo recommends but does not enforce — attribution resolves directly. If visitorId is an internal opaque identifier, a secondary metadata lookup is required, introducing a dependency on the customer's data architecture. NPS Score is only available for customers using Pendo's in-app NPS collection; teams using external survey tools will find this field empty. These integration KPIs reflect the depth and consistency of the team's Pendo instrumentation as much as they reflect actual product performance.