Optimize Customer Journey Analytics and Analysis Workspace performance

Factor

Definition

Influenced by

Optimization

The total number of Freeform table cells in the project, calculated by rows * columns across all tables. Excludes hidden data sources. The guideline is 4000.

Reduce the number of columns in your table to only the most relevant data points. Reduce the number of rows in your table by adjusting the number of rows shown, applying a table filter, or applying a filter.

The total number of components used in the project. The guideline is 100.

The number of used components is not a direct influencer of performance. However, the complexity of those components will contribute to performance of the project. See optimizations in the “Additional factors” section below.

This factor displays the longest date range used the project. The guideline is 1 year.

Where possible, don’t pull in more data than you need. Narrow the panel calendar to the relevant dates for your analysis or use date range components (purple components) in your freeform tables. Date ranges used in a table override the panel date range. For example, you can add last month, last week and yesterday to the table columns to request those specific ranges of data. For more information on working with date ranges in Analysis Workspace, watch this video.

Additionally, minimize the number of year-over-year comparisons used in the project. When a year-over-year comparison is calculated, it looks across the full 13 months of data between the months of interest. This has the same impact as changing the panel date range to last 13 months.

Intricate filters can have a significant impact on project performance.

Factors that add complexity to a filter (in descending order of impact) include:

• Operators of “contains,”, “contains any of”, “matches,” “starts with,” or “ends with”
• Sequential filtering, especially when dimension restrictions (Within/After) are used
• Number of unique dimension items within dimensions used in the filter (e.g., Page = ‘A’ when Page has 10 unique items will be faster than Page = ‘A’ when Page has 100000 unique items)
• Number of different dimensions used (e.g., Page = ‘Home’ and Page = ‘Search results’ will be faster than eVar 1 = ‘red’ and eVar 2 = ‘blue’)
• Many OR operators (instead of AND)
• Nested containers that vary in scope (e.g., “Event” inside of “Session” inside of “Person”)

While some of the complexity factors cannot be prevented, look for opportunities to reduce the complexity of your filters. In general, the more specific you can be with your filter criteria, the better. For example:

• With containers, using a single container at the top of the filter is faster than a series of nested containers.
• With operators, “equals” is faster than “contains”, and “equals any of” is faster than “contains any of”.
• With many criteria, AND operators is faster than a series of OR operators.

Look for opportunities to reduce many OR statements into a single “equals any of” statement.

Visualization complexity (filters, metrics, filters)

The type of visualization (e.g. fallout vs a freeform table) added to a project by itself does not influence project performance very much. It is the complexity of the visualization that adds to processing time.

Factors that add complexity to a visualization include:

• Range of data requested
• Number of filters applied; for instance, filters used as rows of a freeform table
• Use of complex filters
• Static item rows or columns in freeform tables
• Filters applied to rows in freeform tables
• Number of metrics included, especially calculated metrics that use filters

The amount of reporting capacity you and other customers share within an Adobe data center.

This is impacted by the number of concurrent queries made by your organization and other organizations within your data center.

Your organization is entitled to a set capacity and if the system is under a light load, Adobe will shift more capacity to you, above and beyond your entitled allowance.

The number of queries that are being requested by your organization at the same time. Each organization is entitled to a minimum of 5 concurrent queries. If a report is taking a long time, typically it is due to the fact that it is in a queue with other reports. This means your organization is trying to run many concurrent requests against a specific data view.

Queries can come from API requests, reporting UIs (Analysis Workspace, Report Builder, etc.), scheduled projects, scheduled alerts, and concurrent users making reporting requests.

Spread your requests and schedules for the data view more evenly throughout the day. Also, shift your requests to off-peak times when possible. Monday mornings, Tuesday mornings, and the first of each month are peak reporting times.

The amount of data collected into your Connection.

Consult with your implementation team or Customer Journey Analytics expert to determine if there are implementation improvements that can be made to improve overall experience in Customer Journey Analytics.

Highly complex dimensions can have a significant impact on project performance, specifically dimensions or metrics based on complex custom fields.

Reduce the number of custom fields or create separate dimensions.

Also known as high-cardinality dimensions, these dimensions may impact reporting performance.

See high-cardinality dimensions

See high-cardinality dimensions

Factor

Definition

Influenced by

Optimization

Adobe sends in 10 test calls when the performance page is opened. This represents the percentage of those calls to Adobe that succeed.

Local network issues or Adobe issues will impact this factor.

Check status.adobe.com to verify if there are any known service issues. Then, validate your local network connection.

Available for Google Chrome only. Your browser’s estimate of the bandwidth at your location. The guideline is 2.0MB/s.

Your local network connection will impact this factor.

Validate your local network connection.

Adobe sends in 10 test calls when the performance page is opened. This represents the amount of time it takes on average for each request to go to Adobe and be returned. Put more simply, it is a measure of how fast the internet is between your location and Adobe. The guideline is < 1 second.

Local network issues, many open browser tabs, or Adobe issues will impact this factor.

Check status.adobe.com to verify if there are any known service issues. Then, validate your local network connection and close unused browser tabs.

Factor

Definition

Influenced by

Optimization

How fast your computer performs a processing test. The guideline is < 750ms.

Your hardware as well as concurrent programs will impact this factor.

Open your computer’s Task Manager (PC) or Activity Monitor (Mac) to determine if any programs can be closed. Then, close unused browser tabs or other programs.

If those actions do not help, discuss hardware details with your IT team.

Available for Google Chrome only. Every Workspace tab in a Google Chrome browser shares 4GB of memory in total. This represents the percent of that memory allowance that is being consumed by the current project. The guideline is 3500 MB, which is the point at which Workspace will begin surfacing memory errors.

Working in multiple tabs or downloading 50000 rows of data will contribute to increased memory usage.

If you receive a memory error, close other Workspace tabs and/or run 50000 row downloads one-at-a-time.

Data stored locally to your computer for use in the browser. Each origin (e.g. experience.adobe.com) has an allowance of 10MB.

Analysis Workspace uses local storage for several functions, including to store auto-saved (existing) projects, user settings, and feature flags.

To ensure Analysis Workspace functions do not become disrupted, clear local storage for the experience.adobe.com domain.

FPS stands for Frames per second, which is how many times per second the browser draws the page on your screen. 24 FPS is commonly what the human eye can observe; if FPS is lower than that, you will observe rendering issues in Workspace.

FPS is impacted by multitasking across many Workspace projects at once and size of the project being viewed. Other programs running on your computer may have an impact, such as streaming, background scanners, etc. Additionally, your hardware will impact this factor.

Open your computer’s Task Manager (PC) or Activity Monitor (Mac) to determine if any programs can be closed. Then, close unused browser tabs or other programs.

If those actions do not help, discuss hardware details with your IT team.

Factor

Definition

Optimization

The total number of queries (requests) made to Adobe to retrieve data that is displayed in the project. Queries include ranked requests for tables, anomaly detection, sparklines, components shown in the left rail, and more. Excludes collapsed panels and visualizations. The guideline is 100.

Simplify your project where possible by splitting out data into several projects that serve a specific purpose or group of stakeholders. Use tags to organize projects into themes, and use direct linking to create an internal table of contents so that stakeholders can more easily find what they need.

The number of expanded panels out of the total number of panels in the project. The guideline is 5.

After taking steps to simplify your project, collapse panels in your project that do not need to viewed on load. When the project is opened, only expanded panels will be processed. Collapsed panels will not be processed until the user expands them.

The number of expanded tables and visualizations out of the total in the project, including hidden data sources. The guideline is 15.

After taking steps to simplify your project, collapse visualizations in your project that do not need to be viewed on load. Prioritize the visuals that are most important to the consumer of the report and break out supporting visuals into a separate, more detailed panel or project if needed.

See the “Query factors” table above.

See the “Query factors” table above.

See the “Query factors” table above.

Factor

Definition

Optimization

The time required from when the request is initiated to when it is complete. The guideline is 15 seconds.

In the Request processing diagram above, the request time represents the full process, from Analysis Workspace request initiated to Analysis Workspace request complete.

The time required from when the request is initiated to when it is complete.

In the Request processing diagram above, the request time represents the full process, from Analysis Workspace request initiated to Analysis Workspace request complete.

Because Analysis Workspace stores only the hash for any strings that are used in any segments, each time you process a project, Lookups are performed to match the hashes with the appropriate values. The guideline is under 2 seconds.

This can be a resource-intensive process, depending on the number of values that could potentially match the hash.

In the Request processing diagram above, the lookup time is represented in the Lookups phase (at the time of Request Engine processing phase).

If requests are slowing down here, it is probably due to having too many string segments in you project, or having strings with overly generic values that have too many potential matches.

The total time waiting in queue before requests are processed. The guideline is 5 seconds.

In the Request processing diagram above, the queue time is represented in the Request Engine queue phase and Server queue phase.

If requests are slowing down here, it may be due to too many requests running simultaneously in your organization. Try running the request at an off-peak time.

The average amount of time it takes to process the request.

In the Request processing diagram above, the average server processing time is represented in the Server queue phase and Server processing phase. The guideline is 10 seconds

If requests are slowing down here, it is likely that the project has overly long date ranges or complex visualizations. Try shortening your project date range in order to decrease processing times.

Not all requests require the same amount of time to process. Request complexity can help provide a general idea about the time required to process the request. The guideline is Medium or lower.

Possible values include:

• Low
• Medium
• High

This value is influenced by the values in the following columns:

• Month boundaries
• Columns
• Segments

The number of months that are included in a request. More month boundaries adds to the complexity of the request. The guideline is 6 or fewer.

If requests are slowing down here, it may be because the month boundaries in your project are too large. Try reducing the number of months.

The number of metrics and breakdowns in the request. More columns adds to the complexity of the request. The guideline is 10 or fewer.

If requests are slowing down here, it may be because there are too many columns in your project. Try reducing the number of columns.

The number of segments applied to the request. More segments adds to the complexity of the request. The guideline is 5 or fewer.

If requests are slowing down here, it may be because there are too many segments in your project. Try reducing the number of segments.