Adobe Analytics Business Practitioner Expert cheat sheet

A KPI must be specific, measurable, and tied to a named report or metric; vague goals like 'improve engagement' fail the requirement test.

Business requirements map to solution design via a traceability matrix: objective -> KPI -> variable/event -> report.

Success metrics should be defined before implementation; retrofitting KPIs after data collection usually requires rework.

Stakeholder interviews should capture reporting cadence (real-time vs. daily vs. weekly) because it drives data latency and processing choices.

Scope documentation must list in-scope report suites, domains, apps, and out-of-scope items to prevent scope creep.

A use-case inventory (who, what decision, what report) is the primary input to report suite and variable design.

Acceptance criteria for an implementation are testable statements such as 'order confirmation page fires purchase event with revenue'.

Segment and calculated metric requests are requirements too; capture definitions (inclusion rules, numerator/denominator) in the BRD.

Report suite settings control base currency, time zone, default landing page, and whether the suite is a rollup or virtual suite.

Traffic variables (props) are limited to 100 per report suite and accept only one value per hit unless list props are enabled.

Conversion variables (eVars) support allocation and expiration settings; expiration options include visit, page view, and event-based.

Success events can be counter, currency, or numeric; currency events require a currency code and are summed, not counted.

Processing rules run at hit time and can set, copy, or delete values without changing page code; they execute in listed order.

The Adobe Experience Platform Web SDK sends data via a single edge call; the legacy AppMeasurement library uses tracking server calls.

Context data variables (c.) must be mapped to props/eVars via processing rules or the Analytics extension to appear in reports.

Do Not Track and opt-out settings affect data collection; confirm consent management integration before launch.

A segment is a container of hit, visit, or visitor-level rules; the container level determines whether the segment includes whole visits or only matching hits.

Segment stacking is an AND of containers; sequential segments require ordered hits and can include 'then' and 'within' constraints.

Calculated metrics combine existing metrics with +, -, *, / and functions; the formula must reference metrics available in the same report suite context.

Division by zero in a calculated metric returns no data rather than an error; guard with a conditional or choose a nonzero denominator.

Virtual report suites apply a segment and optional permissions, time zone, and currency without copying data; they are not separate data collections.

A virtual report suite inherits its parent's variables and events; you cannot add new eVars or events inside a virtual suite.

Segment and calculated metric sharing is governed by owner and 'shared with' settings; unshared items are visible only to the owner and admins.

Attribution on a segment applies to the segment's inclusion, not to the metric; metric-level attribution is set in the calculated metric or column settings.

Freeform tables support breakdowns, segment drop zones, and column-level settings such as percent, anomaly detection, and attribution models.

The attribution panel compares models (last touch, first touch, linear, participation, same touch) for a chosen conversion event.

Fallout reports show the percentage of visitors who proceed through an ordered set of pages or events; each step can be a segment.

Flow reports visualize pathing before and after a node; the node can be a page, dimension value, or segment.

Anomaly detection flags values outside a statistical confidence band and requires sufficient historical data to train.

Project sharing options are edit, duplicate, and view-only; scheduled projects deliver PDF or CSV to named recipients.

The calendar and date comparison features let you compare current period to prior period, same period last year, or a custom range.

Data warehouse requests and report builder are separate delivery mechanisms from Workspace projects and have their own scheduling limits.

Marketing channel processing rules evaluate in order; the first matching rule wins, so place specific rules above generic ones.

Channel expiration (visitor vs. visit) determines how long a channel value persists before being overwritten by a new touch.

The 'Internal' and 'Direct' channels are default fallbacks; unclassified traffic lands in 'No Channel' or 'Unspecified' depending on configuration.

Attribution IQ models include last touch, first touch, linear, participation, same touch, U-shaped, J-shaped, and time decay; model choice changes credit, not raw hit counts.

Cross-device analytics requires an identity graph or customer ID; without it, visitor counts are device-based and overstate unique people.

Data governance covers classification (e.g., PII flags), data retention, and access controls; classification labels drive downstream privacy rules.

IP obfuscation and geolocation settings are report-suite level and affect both reporting accuracy and privacy compliance.

Data repair and data insertion are limited admin tools; document changes because they alter historical reporting and are not reversible without a backup.

Product profiles in the Admin Console grant Analytics permissions; group membership determines report suite access and tool rights.

Report suite access can be 'all', 'none', or a specific list; a user without access sees no data even if the login succeeds.

The Analytics admin console includes user management, report suite settings, data sources, and classification importers.

Common data discrepancy causes: different date ranges, time zone settings, segment definitions, bot filtering, and attribution differences.

The Adobe Experience Cloud debugger and browser developer tools confirm whether a tracking call fires and which variables are set.

Processing rules and VISTA rules can silently alter or drop data; audit the rule list when values are missing.

Hit-level data latency is typically minutes for real-time reports and longer for full processing; check processing status before escalating.

Alerts and scheduled reports fail silently if the owner loses access; verify ownership and permissions when deliveries stop.