Analytics Dashboards

GermainUX Analytics Dashboards help business and technical teams detect, analyze, and resolve the issues affecting application adoption, eCommerce conversion,decima employee productivity, user experience, and technology performance.

They connect business outcomes with user behavior, workflows, application performance, errors, and individual user sessions. This enables teams to answer five essential questions:

GermainUX provides preconfigured dashboards for common use cases. Every dashboard can also be customized with the KPIs, portlets, measures, pivots, filters, time ranges, permissions, and visualizations required by each team.

🎚️ Four Levels of Analysis

GermainUX provides four connected levels of visualization and analysis:

Users can move from a high-level business or technology symptom to the individual event, request, trace, or user session responsible for it. Filters and visibility are preserved throughout the investigation.

📑 Business and Technical KPIs

Analytics Dashboards support two categories of KPIs:

An administrator can assign users to Business or Technical teams so they see the dashboards and KPIs most relevant to their responsibilities.

🧭 Aggregate Dashboard

The Aggregate Dashboard is the top-level view and the starting point for most investigations. It consolidates business and technical KPIs so teams can monitor overall health, identify trends, quantify business impact, and determine which issues require deeper analysis.

The aggregation frequency is configurable.

Add an Aggregate Dashboard

Go to:

Germain Workspace > Dashboards > All > +

Then:

🔎 Filters

Filters can be applied at two levels:

Use dashboard-level filters when the same condition should apply to the entire dashboard. This is usually easier to manage and provides a consistent investigation context.

🔒 Permissions

🔁 Compare To

Compare To makes it easier to identify increases, decreases, and regressions by comparing the selected period with another period. It is available at the dashboard level across Aggregate, Pivot, Drill-through, and Analysis dashboards.

Comparisons display:

Use Compare To to distinguish a genuine regression from normal variation or seasonality.

📌 Pin a Dashboard

Pin frequently used Aggregate, Pivot, or Drill-through dashboards to make them directly accessible from the Workspace menu.

Pinned dashboards can be reordered under:

Workspace Menu > Dashboards

🔢 Decimal Precision

A default decimal precision can be configured at the Workspace level. It can be overridden for an individual dashboard or portlet when a different level of precision is required.

📋 Copy to Clipboard

Values and detailed insights can be copied to the clipboard for use in tickets, reports, messages, and other investigation tools.

🧩 Portlets

Portlets are configurable widgets used to visualize KPI data.

Portlet	Purpose
General Health	Displays trends, distributions, or aggregated KPI health.
Tabular	Displays aggregated data broken down by one or more dimensions.
N-Tabular	Displays several KPIs in a single table.
Map	Displays geographic data. Google web-service access must be permitted by the network configuration.
Flow	Visualizes customer journeys, workflows, or click paths.
Focused Flow	Breaks down journey milestones to identify friction and abandonment at scale.
Funnel	Visualizes the progression and drop-off across a business process or workflow.
Counter	Displays a single KPI value or count.
Scripted	Displays many KPI value or count.

🔍 Focused Flow

Focused Flow analyzes user behavior across journey milestones so teams can find the friction points affecting users at scale, without replaying every session individually.

Supported KPI types include:

• Shopify Events

• Story Beats

• User Clicks

Available options may vary according to the monitored application and KPI.

🖊️ Add or Edit a Portlet

To add a portlet, click + Add Metric on an Aggregate Dashboard. To modify an existing portlet, open its edit settings.

Portlet settings can include:

🚫 Portlet Constraints

Constraints permanently restrict the data included in a portlet. A basic constraint compares a model attribute with a value:

(user.name == 'jsmith')

Use || to include alternative values:

(user.name == 'jsmith') || (user.name == 'bjones')

Use && to combine dimensions:

((user.name == 'jsmith') || (user.name == 'bjones')) &&
(application.name == 'Siebel')

⏳ Portlet Filters

Quick filters can be applied to KPI measures or text using operators such as:

For example, a filter containing Error% can focus the visualization on error-related values. For a permanent filter, edit the portlet and save the filter or constraint in its configuration.

🧮 Measures and Pivots

Measures determine which numeric value is analyzed, such as count, average duration, total lost productivity, or conversion rate.

Pivots break a KPI down by dimensions such as:

Configured pivots in a Tabular Portlet can be reordered by dragging their column headers.

🏷️ Display Names

The display names of KPI measures and pivots can be changed under:

Germain Workspace > Analytics > KPI

For example, Total Feedback Comment # can be displayed as Real User Opinion Count.

⚙️ Manage an Aggregate Dashboard

Available management actions include:

• Edit

• Clone

• Print

• Change trend granularity

• Create a note or ticket

📁 Preconfigured Dashboards

GermainUX includes preconfigured Aggregate Dashboards, along with dashboards created by your organization. They are available under:

Germain Workspace > Dashboards > All

✨ Example of How to Create an Aggregate Dashboard

🔁 Pivot Dashboard — Second Level

The Pivot Dashboard breaks down the aggregated data selected from an Aggregate Dashboard. It helps teams determine where a KPI or issue is concentrated.

For example, a conversion decline can be pivoted by product, journey step, region, device, or customer segment. A CRM performance problem can be pivoted by page, action, user group, browser, or application version.

Pivot Dashboards:

• Are generated automatically from Aggregate Dashboards

• Inherit filters and visibility from the originating dashboard

• Can be further edited when a different analysis is required

• Provide drilldowns to Drill-through and Analysis dashboards

To create or modify a Pivot Dashboard, click + from the Pivot Dashboard and configure the required portlets, measures, and dimensions.

🚦 Drill-through Dashboard — Third Level

The Drill-through Dashboard displays the individual data points behind an Aggregate or Pivot value. It answers questions such as:

• Which transactions failed?

• Which users experienced the issue?

• Which sessions abandoned the journey?

• Which requests were slow?

• Which workflow instances exceeded their target duration?

Open Drill-through from an Aggregate or Pivot Dashboard using the list or binoculars action.

Drill-through Dashboards:

• Display raw instance-level data

• Inherit filters and visibility from the originating dashboard

• Allow columns to be added, removed, or reordered

• Provide a direct drilldown to the Analysis Dashboard

Available actions include:

• Refresh

• Edit

• Export to CSV

• Print

• Add a note or ticket
  

🔬 Analysis Dashboard — Fourth Level

The Analysis Dashboard is GermainUX's root-cause investigation screen. It helps investigators move from a KPI symptom to an evidence-backed root cause by combining baselines, relationships, related KPIs, segmentation, individual instances, technical traces, and Session Replay in one workflow.

It takes a KPI such as a slow click, a conversion drop, an increase in errors, or a stalled workflow and answers three questions:

1. Is this actually abnormal? Compared with what, and by how much?

2. What else changed at the same time? Which related signals moved with it?

3. Where is the issue concentrated? Which users, pages, products, processes, regions, devices, or versions are affected?

The Analysis Dashboard can be opened from Aggregate, Pivot, or Drill-through dashboards. It inherits the current time range, filters, and visibility so the investigation remains in context.

❓ What Can Be Diagnosed?

The same Analysis workflow supports three major diagnostic categories:

These diagnoses help organizations:

• Improve CRM, ERP, and internal-application adoption

• Increase eCommerce conversion

• Improve employee productivity

• Reduce the time required to diagnose and resolve digital issues

🗂️ Aggregate and Instance Analysis

🚀 Why Use the Analysis Dashboard?

Analysis replaces manual spreadsheet pivots and tool-hopping with a guided investigation workflow.

• Automatic context: Compare the KPI with recent baselines, previous periods, trends, and its statistical distribution.

• Precomputed leads: Surface related KPIs and leading segments instead of requiring investigators to guess where to look.

• Evidence-based drilldown: Move from an aggregate trend to individual instances, technical traces, HAR exports, and Session Replay.

• Business-impact prioritization: Focus corrective action on the journeys, screens, products, processes, and user groups with the greatest impact.

It is designed for support teams, operations, developers, SREs, product teams, UX teams, business-process owners, conversion teams, and analysts.

🏗️ KPIs: The Building Blocks of Analysis

Every Analysis Dashboard is built from KPIs collected by GermainUX. A KPI can represent a click, request, error, rage click, conversion, workflow step, deployment, system metric, alert, or other measurable fact.

Each KPI can define:

Hover over a KPI badge to see its definition and thresholds. These thresholds determine the status and coloring shown throughout the analysis.

📈 Aggregate Analysis

Aggregate Analysis contains four primary sections.

🔖 1. Summary

The Summary establishes whether the KPI is abnormal and how significant the change is.

• Vs Recent: Compares the current value with a rolling recent baseline. Details show the current value, baseline, contributing periods, and instance counts.

• Vs Last Quarter: Compares the current period with the corresponding period one quarter earlier, helping distinguish regressions from seasonality.

• Trend: Displays how the KPI changed during the selected range. Granularity adapts to the time range.

• How Extreme: Positions the selected value within the KPI distribution, helping determine whether the problem affects the overall population or only a slow or problematic tail.

A shifted median usually indicates broad impact. A high p95 or p99 with a normal median usually indicates that a subset of users or instances is affected.

🔗 2. Relationships

The Relationships chart overlays the selected KPI with related signals on a shared timeline.

It can display:

• The selected KPI

• The overall population as a reference

• Related continuous KPIs as lines

• Discrete events, errors, or alerts as markers

• Dimensional pivots of a related KPI

Use this view to confirm whether an error increase, request spike, deployment, user action, or other signal occurred at the same time as the symptom.

⛓️ 3. Related KPIs

Related KPIs are ranked according to how they changed alongside the KPI being investigated. Each related KPI can include:

• A SPIKE or DROP indicator

• Current value versus baseline

• Absolute or percentage variation

• Trend and baseline visualization

• Co-movement score

The co-movement score helps distinguish a signal that consistently moves with the investigated KPI from unrelated background noise. When choosing which leads to investigate, prioritize meaningful co-movement over a large delta alone.

🧭 4. Segmentation and Leading Factors

Segmentation scans the available KPI dimensions and surfaces the segments most likely to explain the change.

Dimensions adapt to the KPI. Examples include:

• A Salesforce click: name, page title, page path, target, user, browser, or application version

• An eCommerce conversion: product, variant, category, journey step, region, device, or customer segment

• A business workflow: process, milestone, owner, department, status, or outcome

Each segment can show its value, trend, change versus baseline, and counts of instances, sessions, and users. These counts help distinguish an isolated problem from a systemic issue.

👥 Instance Analysis

Opening an individual instance reframes the investigation around two questions:

1. Why did this specific instance behave this way?

2. How far does the same problem extend?

⚖️ Comparison Scopes

The scope selector compares the selected instance at five levels:

INSTANCE → SESSION → USER → POPULATION → OVERALL

• Instance: The selected event or transaction

• Session: Comparable events within the same user session

• User: Comparable events for the same user

• Population: Comparable events sharing the KPI's configured matching fields

• Overall: All instances within the current investigation context

Reading these scopes from left to right quickly shows whether the issue is isolated, user-specific, limited to a particular population, or systemic.

The MATCHING · N indicator shows how many instances are considered comparable. Matching is determined by the KPI's configured SLA Default Value fields.

⏱️ Duration Breakdown

For duration-based KPIs, the Analysis Dashboard can divide elapsed time into:

• Service

• Network

• Rendering

• Document Processing

• Idle or Unmonitored

Duration is available in two forms:

• Exclusive: Time spent in the selected span only

• Cumulative: Time spent in the selected span and its child spans

This helps determine whether the bottleneck is associated with a backend service, network, browser rendering, document processing, or unmonitored activity.

🆔 Attributes and Identifiers

Instance details can include:

• Application and environment

• User and session

• Page, screen, target, or business object

• Host and endpoint

• Heap or resource usage

• Lost productivity

• First Input Delay, Largest Contentful Paint, and long-task duration

Copyable identifiers can include:

• Instance ID

• Correlation ID

• Session ID

• Sequence

• Tab ID

Use these identifiers to connect GermainUX evidence with application logs, APM traces, support tickets, or other systems.

🕰️ Trace Timeline

The Trace Timeline displays the selected transaction, child spans, server timings, and surrounding browser or application events on a shared time axis.

Selecting an HTTP span can display:

• DNS, connection, SSL, stalled, wait, and download time

• Response size

• Cache and compression information

• Full request URL

• Response details, where captured and permitted

• HAR export
  

  

✨ Extended Context

The Trace Timeline can be widened beyond the selected transaction to show what happened before and after it, including:

• Console errors

• User-facing errors

• Rage clicks

• Popups

• Navigations

• Related requests and transactions

The cause of a symptom may occur outside the duration of the transaction that first surfaced it.

🔁 Recursive Analysis

The Analysis → action on a trace span opens a new Analysis Dashboard for that span's KPI. This supports investigations such as:

Slow user click → Slow HTTP request → Slow service call → Slow database operation

Each level receives its own baselines, relationships, segmentation, and individual evidence.

📋 Standard Investigation Workflow

1️⃣ 1. Start from a symptom

Open Analysis from a dashboard KPI, SLA violation, friction, alert, or Drill-through Dashboard.

2️⃣ 2. Confirm that the issue is real

Review Vs Recent, Vs Last Quarter, Trend, and How Extreme. Determine whether the signal represents a regression, seasonality, a broad shift, or a problematic subset.

3️⃣ 3. Identify what changed with it

Review Related KPIs. Prioritize signals with meaningful co-movement and overlay the most promising candidates on the Relationships chart to verify that the timing aligns.

4️⃣ 4. Find where the issue is concentrated

Review Leading Factors. Prioritize DEGRADED and NEW segments with significant user, session, or instance reach.

A NEW segment appearing at the same time as a regression may indicate a deployment, release, configuration change, or new routing path.

5️⃣ 5. Obtain instance-level evidence

Open a representative instance from an affected segment. Use the scope comparison to determine the reach, the duration breakdown to locate the bottleneck, and the Trace Timeline to identify the responsible request or operation.

6️⃣ 6. Confirm user impact

Widen the trace window and inspect surrounding errors and user behavior. Open Session Replay where available to see how the issue affected the user's experience.

Use Previous and Next to compare neighboring instances and confirm that the pattern is repeatable.

7️⃣ 7. Share the finding

Copy identifiers, export a HAR or CSV where appropriate, print or share the analysis, and attach a note or ticket. The resulting evidence chain connects business impact to the underlying technical, workflow, or UX cause.

⏩ Reading the Analysis Quickly

• Status colors are based on the KPI's configured SLA thresholds.

• Instance, session, and user counts indicate the reach of a segment.

• n/a for a baseline delta usually means there is no comparable history and the segment may be NEW.

• Co-movement is generally more useful than raw variation when choosing which related KPI to investigate.

• A retention warning means aggregate data remains available, but the raw instances or traces for that period have been purged.

🔧 Example: Diagnosing Slow Salesforce Clicks

Assume the Salesforce User Click KPI averages 5.175 seconds over seven days. The overall average appears to be improving compared with its recent baseline, but the p95 is 10.913 seconds.

This indicates a subset problem rather than a broad slowdown.

Related KPIs show that JavaScript console events and outbound HTTP requests increased during the same period, while user key presses decreased. The Relationships chart confirms that the increases align with the slow-click period.

Leading Factors then identify that the issue is concentrated on Salesforce Lightning record-view pages. One page path is DEGRADED across multiple sessions and users, while another path is marked NEW and DEGRADED, suggesting a recent release or routing change.

Opening a representative instance shows that nearly all its duration is Service time. Its outbound request spends most of its time waiting for the server to begin responding, rather than connecting, downloading, or rendering.

Extending the trace window reveals user-facing errors and rage clicks. Session Replay confirms that users repeatedly click because the interface does not respond.

The final diagnosis is supported by a complete evidence chain:

Salesforce Lightning record-view pages became slower following a recent change. Increased wait time on the related backend endpoint produced user-facing errors and rage clicks across the affected user population.

🛒 Example: Diagnosing an eCommerce Conversion Issue

For an eCommerce conversion KPI, the same workflow can compare the current conversion rate with historical baselines, surface related changes in errors or page performance, and identify the products, variants, journey steps, devices, regions, or customer segments most affected.

The diagnostic method remains consistent, but the dimensions and outcome are commercial rather than technical. Teams can move from a conversion decline to the affected journey segment, then to individual customer sessions and the underlying UX or technology friction.

🔧 Analysis Configuration

Analysis configuration is available under:

Settings > Analytics

⚙️ KPI Configuration

An SLA can use expressions such as:

Threshold Failed: (success == false) && (duration >= 5)
Threshold: duration >= 5

Advanced KPI settings expose the monitoring components, agents, and data-source configurations that feed the KPI.

The SLA Default Value fields determine which instances are treated as equivalent for instance-level comparisons. Configure these fields carefully to ensure comparisons use genuinely comparable instances.

🔗 Relationship Configuration

Relationships can be configured for a pair of KPIs.

Well-configured relationships make Related KPIs more relevant and reduce investigation noise.

🛠️ Supporting Configuration

⌛ Retention

Aggregate data usually remains available longer than raw instance data. When raw data has been purged, aggregate analysis and segmentation can still be available, but individual traces and instance-level drilldowns are not.

Configure retention according to the organization's realistic investigation window. If issues are normally reported several weeks after they occur, raw-data retention should be long enough to preserve the evidence required for diagnosis.

🤖 Advanced Analysis Capabilities

• AI Explore: Ask questions about the analysis conversationally.

• Dimensional pivoting: Split a related KPI into its leading dimension values and overlay them on the Relationships chart.

• Recursive analysis: Move from a user action to its requests, services, and other contributing operations.

• Extended trace windows: Inspect events before and after the selected transaction.

• Instance navigation: Use Previous and Next to compare instances without returning to Drill-through.

• Correlation identifiers: Connect GermainUX evidence with logs, traces, and tickets.

• HAR export: Export HTTP details for offline investigation or external teams.

• Session Replay: Visually confirm the user's experience and behavior.

• Trace filtering: Focus the timeline on relevant KPIs and hide unrelated activity.

• Cross-scope comparison: Determine whether an issue is isolated, user-specific, population-specific, or systemic.

• Notes and annotations: Keep findings attached to the underlying evidence.

• Adaptive time ranges: Expand or narrow the investigation period while GermainUX adjusts the trend granularity.

📂 Other Analytics Dashboards

🖥️ Operational Dashboard

The Operational Dashboard provides a 360-degree view of process, user, and technology performance. It gives business and IT operations teams a consolidated view of the current state of their applications, workflows, technologies, and users.

Data is aggregated across KPIs and supplied by the GermainUX real-time data engine. The dashboard is customizable, including the columns displayed across its tabs.

Friction Dashboard

The Friction Dashboard provides a consolidated view of issues detected and analyzed by GermainUX. Use it to:

• Review detected UX, workflow, and technology frictions

• Understand their reach and business impact

• Prioritize the most consequential issues

• Open Analysis to identify their root causes

• Track investigation and resolution

💡 AI Advice Dashboard

The AI Advice Dashboard presents prioritized recommendations for addressing the frictions identified by GermainUX. It helps teams focus on the actions most likely to improve adoption, conversion, productivity, user experience, or application performance.

📋 SLA Violation Dashboard

The SLA Violation Dashboard, called the Insights Dashboard in older versions, displays SLA violations by severity, priority, maximum variance, and other dimensions.

Maximum Variance helps teams identify the largest regressions and newly emerging issues.

🔧 Actions Available from a KPI

Hover over a KPI on an Aggregate Dashboard to access the following actions.

🔎 Analysis

Opens the Analysis Dashboard for the selected KPI, preserving the current filters and time range.

💀 Worst Instance

Opens Analysis for the worst individual data point in the current context. GermainUX calculates an instance score using configurable negative factors and selects the instance with the highest score.

For User Session Replay, a score can incorporate factors such as:

• User-facing errors

• Rage clicks

• Negative feedback

• Searches with no results

• Dead clicks

• Click-initiated errors

The scoring formula and factor weights are configurable. See Score Configuration.

📑 List All Instances

Opens the Drill-through Dashboard containing all individual data points for the selected KPI and current context.

📊 Pivot by Dimensions

Breaks down the selected KPI by its configured dimensions to identify the values contributing most to the result.

👁️ Watch This Metric

Creates a notification for a selected KPI and condition. See Watch Notifications.