While Adobe Analytics is an enterprise tool designed to handle many users pulling detailed reports, performance can degrade if enough complex simultaneous report requests are made. Most performance issues are compounded by multiple factors. Adobe recommends adhering to the following general guidelines:
The following table outlines the hard limits of each Analytics capability:
While several capabilities do not impose a limit on some features, there is still a reasonable limit that can be reached. This limit depends on the amount of data and complexity of reporting. Complex reporting reaches those soft limits faster than simple reports.
- Row limit: While the current maximum of 400 rows per visualization is in place, keeping it smaller (10-50 rows) optimizes performance. If more than 400 rows is needed, using an alternate tool like Data Warehouse is recommended. Generally speaking, the fewer rows a visualization has, the better it performs. For example, a monthly granular report performs better than a daily granular report, given the same date range.
- Visualization/panel limit: Currently there is not a limit on the number of visualizations an individual project can have. The recommended limit from Adobe is approximately 20 separate visualizations per project. More can be used for simple requests, such as a single day’s worth of data. Less would be advised in larger or more complex reports, such as a daily trend spanning more than a year. Splitting visualizations into multiple projects is recommended if a project starts getting too many panels.
- Number of metrics in a single panel: Adobe Analytics offers the ability to add virtually as many metrics to a freeform table as your screen can fit. In addition to readability issues, performance issues can result if more than 30 metrics are added to a single panel.
- Number of breakdowns in a single panel: Consecutive breakdowns, especially applied to many line items, adds up quickly. While no limit is imposed, Adobe recommends 2-3 breakdowns per panel. More can be used for relatively simple reports.
- Contribution Analysis and Segment Comparison: These two features within Analysis Workspace are powerful, but also resource-intensive. Adobe recommends limiting the number of these reports to 100 per day.
- Search filters: For ranked reports using a search filter, the recommended limit is 50 items.
- Concurrent requests per report suite: Many users querying the same report suite simultaneously can sometimes cause issues, especially for large projects or complex segments. Queries include inside the product and external queries, such as the API. Adobe recommends that users in large organizations be cognizant and courteous to the number/complexity of requests made to the same report suite. If users want to look at data at regular intervals, scheduling reports at non-peak hours is a great way to do so.
- Date ranges: Adobe Analytics is optimized to provide the fastest reporting for the current month. For best performance, use date ranges that avoid spanning month and year boundaries where possible. Long date ranges, especially dates that span year boundaries, are much more processing intensive.
- Number of rules in a segment: The fewer rules a segment has, the better it performs. Adobe Analytics in most browsers start to run into performance issues between 30-50 segment rules. If you have segmentation needs that require looking for many individual line items, using a classification is highly recommended. For example, you want to look at value1, value2, all the way to value1000. Instead of using 1000 segment rules, upload a classification for the variable in question, then use the classification in the segment.
- Upload a classification file to Adobe where value1, value2, through value1000 all contain the same classification value of classified1.
- Once the classification file is processed, create a segment rule where the classification value equals classified1.
- This single segment rule allows you to see all 1000 values in reporting.
- Segment rule logic: The operators 'contains one of' and 'contains all of' are processing intensive, and can cause reports to slow if not used in moderation. These operators are designed to only look at a few values, and not a large complex string. Consider using the 'equals' operator where possible to optimize performance. The 'contains' operator is slower than 'equals', but is faster than 'contains one of' and 'contains all of'.
- Total number of segments in a login company: Segments merely existing don't tend to have a major impact on performance. However, performance issues can result while creating Workspace projects or managing segments if a company has more than 20,000 saved segments.
- Classifications: Classifications are a great way to customize how data is organized and grouped. Classification templates are limited to 30 columns, not including the key value. Using the Classification Rule Builder is highly recommended wherever possible. Only use the Classification Importer when necessary, and only include changed values (instead of all values). See Classification Importer Tips for more information on how to best use classifications to maintain peak performance.
- Number of time periods in a single request: Report Builder imposes a hard limit of 366 time periods in a single request. For example, viewing a daily trend over the past year would be available, but not longer. If more than a year's worth of time is required, use a less granular range, such as weekly or monthly. If more than 366 time periods are required, using Data Warehouse is recommended.
- Number of data requests in a single workbook: Report Builder has a hard limit of 1000 data requests in a single workbook. If you find a report reaching this limit, Adobe recommends using multiple workbooks.
- Number of metrics in a single request: The maximum number of metrics allowed in a single Report Builder or API request is 30.
- Number of rows in a single request: The maximum number of rows allowed in a single request is 50,000. If more than 50,000 rows is needed, use Data Warehouse.