DocumentationAnalyticsComponents Guide

Classification rules

Last update: Thu Feb 02 2023 00:00:00 GMT+0000 (Coordinated Universal Time)

CREATED FOR:

  • User
  • Admin

Classification rules regularly look for unclassified terms. If a rule match is found, the rules add the terms to your classification data tables automatically. You can also use classification rules to overwrite existing keys.

Analytics > Admin > Classification Rule Builder

The Rule Builder lets you create a classification rule set, which is a list of classification rules. A rule matches criteria you specify, then performs an action.

Classification rules are convenient for:

  • Email and Display ads: Create classification rules to group individual display ad campaigns so that you can learn how the Display campaigns are performing against email campaigns.

  • Tracking codes: Create classification rules to categorize key values derived from strings in tracking codes, and match them to specific criteria you define.

  • Search terms: Use regular expressions and wildcards to simplify classifying of search terms. For example if a search term contains baseball, you can set a Sports League classification to MLB.

For example, assume that a tracking code for an email campaign ID is:

em:Summer:2013:Sale.

You can set up three rules in a rule set that identify the parts of the string, then classify the values:

Select Rule Type
Enter Match Criteria
Set Classification
To
Starts With
em:
Channel
Email
Ends With
Sale
Type
Sale
Contains
2013
Year
2013

How Rules Are Processed how-rules-are-processed

Important information about how classification rules are processed.

NOTE
The Rule Builder does not support Numeric 2 classifications.

Important Information about Rules

  • Specify group permissions for classifications in Admin Tools.

  • Regular expressions: Help is available under Regular Expressions in Classification Rules.

  • Report suites: You cannot choose a classification until at least one report suite is selected. You cannot apply the report suite until you have created the rule set and assigned a variable.

    When you test the rule set, use keys (the variable being classified) from the report to see how they will be impacted by the rule set. (The key is the variable being classified, or the first column in the classification upload table.)

  • Rule priority: If a key matches multiple rules that set the same classification (in the Set Classification column), the last rule that matches the classification is used. See About Rule Priority.

  • Limits on number of rules: No set limit exists for the number of rules you can create. However, a large number of rules may impact browser performance.

  • Processing: Rules are processed at frequent intervals, depending on your volume of classification related traffic.

    Active rules process every four hours, examining classification data going back typically one month. The rules automatically check for new values and upload the classifications using the importer.

  • Overwriting existing classifications: See When Do Rules Not Classify Keys? If necessary, you can delete or remove existing classifications, using the importer.

When Do Rules Not Classify Keys?

When you activate rules, you can overwrite existing classifications. In the following situations, a classification rule does not classify a key(variable) if:

  • The key is already classified and you do not select Overwrite Classifications.

    You can overwrite classifications when adding and activating a rule, and when activating a data connectors integration. (For data connectors, rules are created by partners in the Dev Center and displayed in the Classification Rule Builder.)

  • A classified key has not appeared in the data after a time frame specified when overwriting a key, even after you enable Overwrite Classifications.

  • The key is not classified and the key is never passed into Adobe Analytics after the time frame beginning about one month ago.

    note note
    NOTE
    In reports, classifications apply to any time frame specified, whenever a key exists. The date range of a report does not affect reporting.

Regular Expressions in Classification Rules regex-in-classification-rules

Use regular expressions to match consistently formatted string values with a classification. For example, you can create a classification from specific characters in a tracking code. You can match particular characters, words, or patterns of characters.

NOTE
As a best practice, regular expressions are best suited for tracking codes that use delimiters.

Regular Expression - Tracking Code Example section_2EF7951398EB4C2F8E52CEFAB4032669

NOTE
If the tracking code is URL encoded, it will not be classified by the Rules Builder.

In this example, assume you want to classify the following campaign ID:

Sample Key: em:JuneSale:20130601

The parts of the tracking code you want to classify are:

  • em = email
  • JuneSale = campaign name
  • 20130601 = date

Regular Expression: ^(.+)\:(.+)\:(.+)$

How the regular expression correlates to the campaign ID:

Match Groups: Shows how the regular expression corresponds to the campaign ID characters, so that you can classify a position in the campaign ID.

This example tells the rule that the campaign date 20140601 is at the third group (.+), identified by $3.

Rule Builder

In the Rule Builder, configure the rule as follows:

Select Rule Type
Enter Match Criteria
Set Classification
To
Regular Expression
^(.+):(.+):(.+)$
Campaign Date
$3

Syntax

Regular Expression
String or Match Result
Corresponding Match Groups
^(.+)\:(.+)\:(.+)$
em20130601
$0: em20130601 $1: em $2: JuneSale $3: 20130601
Building the syntax
^ = starts the line () = groups characters and lets you extract matching characters in the parentheses. (.+) = captures one ( . ) character and ( + ) any more \ = start of a string. $ = indicates that the preceding character (or character group) is the last in the line.

See Regular Expressions - Reference Table for information about what the characters in a regular expression mean.

Regular Expression - Classifying a Specific Character section_5D300C03FA484BADACBFCA983E738ACF

One way to use a regular expression is to classify a specific character in a string of characters. For example, assume that the following tracking code contains two important characters:

Sample Key: 4s3234

  • 4 = brand name
  • s = identifies a search engine, such as Google

Rule Builder

In the Rule Builder, configure the rule as follows:

Select Rule Type
Enter Match Criteria
Set Classification
To
Regular Expression
^.(s).*$
Brand and Engine
$0 (Captures the first two characters for brand name and search engine.)
Regular Expression
^.(s).*$
Search Engine
$1 (Captures the second character for Google.)

Regular Expressions - Matching Tracking Codes of Varying Length section_E86F5BF5C2F44ABC8FFCE3EA67EE3BB2

This example shows how to identify specific characters between colon delimiters when you have tracking codes of varying lengths. Adobe recommends using one regular expression for each tracking code.

Sample Keys:

  • a:b
  • a:b:c
  • a:b:c:d

Syntax

Rule Builder

In the Rule Builder, configure the rule as follows:

Select Rule Type
Enter Match Criteria
Set Classification
To
Regular Expression For match string a:b
^([^\:]+)\:([^\:]+)$
a
$1
Regular Expression For match string a:b
^([^\:]+)\:([^\:]+)$
b
$2
Regular Expression For match string ac
^([^\:]+)\:([^\:]+)\:([^\:]+)$
a
$1
Regular Expression For match string ac
^([^\:]+)\:([^\:]+)\:([^\:]+)$
b
$2
Regular Expression For match string ac
^([^\:]+)\:([^\:]+)\:([^\:]+)$
c
$3
Regular Expression For match string ac:d
^([^\:]+)\:([^\:]+)\:([^\:]+)\:([^\:])$
d
$4

Regular Expressions - “Does Not Contain” Example section_FCA88A612A4E4B099458E3EF7B60B59C

This example provides a regular expression that matches any string that does not contain specific characters, in this case 13.

Regular expression:

^(?!.*13.*).*$

Test strings:

a:b:
a:b:1313
c:d:xoxo
c:d:yoyo

Match results:

a:b:
c:d:xoxo
c:d:yoyo

In this result, a:b:1313 does not indicate a match.

Regular Expressions - Reference Table section_0211DCB1760042099CCD3ED7A665D716

Expression
Description
(?ms)
Makes the entire regular expression match against a multi-line input, allowing the . wildcard to match any newline characters
(?i)
Makes the entire regular expression case insensitive
[abc]
A single character of: a, b or c
[^abc]
Any single character except: a, b, or c
[a-z]
Any single character in the range a-z
[a-zA-Z]
Any single character in the range a-z or A-Z
^
Start of line (matches the beginning of the line)
$
Match the end of the line (or before newline at the end)
\A
Start of string
\z
End of string
.
Match any character (except a new line)
\s
Any whitespace character
\S
Any non-whitespace character
\d
Any digit
\D
Any non-digit
\w
Any word character (letter, number, underscore)
\W
Any non-word character
\b
Any word boundary
(...)
Capture everything enclosed
(a\b)
a or b
a?
Zero or one of a
a*
Zero or more of a
a+
One or more of a
a{3}
Exactly 3 of a
a{3,}
3 or more of a
a{3,6}
Between 3 and 6 of a

A good resource for testing regular expression validity is https://rubular.com/.

About Rule Priority

If a key is matched to multiple rules, and it sets the same classification column shown in the Set Classification column, the last rule is used. As such, you might want to rank the most important last in your rule set.

If you create multiple rules that do not share the same classification, processing order does not matter.

What follows a search-term rule example that classifies search types for an athlete:

Rule Number
Rule Type
Match
Set Classification
To
1
Contains
Cowboys
Search Type
Team
2
Contains
Fantasy
Search Type
Fantasy
3
Contains
Romo
Search Type
Player

If a user searches for Cowboys fantasy Tony Romo, the term Player is classified, because it matches the last given classification shown in the Set Classification column.

Similarly, suppose you set up two rules in a set for the following search terms:

Rule Number
Rule Type
Match
Set Classification
To
1
Contains
Cowboys
City
Dallas
2
Contains
Broncos
City
Denver

A user searches for Cowboys vs. Broncos. If the rule builder finds a conflict in rule matching, the classification for the second rule (Denver) applies to this search.

Add a Classification Rule to a Rule Set add-classification-to-rule-set

Add rules by matching a condition to a classification, and specifying the action.

NOTE
In this procedure, you must apply the rules to one or more report suites. The recommended number of rules per rule set is between 500 and 1000, although there are no limits. If you have over 100 rules, consider simplifying your rule set by using sub-classifications.

To add or edit a classification rule:

  1. Create a Classification Rule Set .

  2. On the rule set page, click Add Rule.

  3. Next to Report Suites, click Add Suites to specify one or more report suites to assign to this rule set.

    The Select Report Suites page displays.

    note note
    NOTE
    Report suites display on this page only when the following conditions are met:

    • The report suites have at least one classification defined for that variable in Admin Tools.

      (See Variable in Classification Rule Sets for an explanation about this prerequisite.)

    • You selected the report suite on the Available Report Suites page, which displays after you click Add Rule Set to create the rule set.

  4. Specify whether to overwrite existing values:

    table 0-row-2 1-row-2
    Rules overwrite any existing values (Default setting) Always overwrite existing classification keys, including classifications uploaded via the importer (SAINT).
    Rules overwrite only unset values Only fill in blank (unset) cells. Existing classifications will not be changed.

  5. Define the rule or rules.

    Step Result

    For examples of building rules, see Classifications Rule Builder and Regular Expressions in Classification Rules.

    note note
    NOTE
    If a key matches multiple rules that set the same classification (in the Set Classification column), the last rule that matches the classification is used. See About Rule Priority above for more information about sorting rules.

  6. Test your rule set.

  7. After testing, click Active to validate and activate the rule.

    Activating a rule automatically builds the file and uploads it for you.

    Field definitions: See Classification Rule Builder for complete definitions of interface options on this page.

Test a Classification Rule Set

You can test a classification rule or rule set. Running a test checks all the rules in a set.

To test a Classification Rule Set:

  1. Create a Classification Rule Set .

  2. On the Classification Rule Builder, click the rule set name.

  3. Ensure that the rule set is associated with a report suite.

  4. On the rule editor, click Test Rule Set.

    Step Result

  5. Type or paste test keys in the Sample Keys field.

    Sample keys include:

    • Tracking codes
    • Search keywords or phrases

    See Regular Expressions in Classification Rules for information about testing regular expressions.

  6. Click Run Test.

    Rules that match are displayed in the Results table.

  7. (Optional) Click Activate to activate the rule, and to overwrite existing classifications.

    See for more information about using rules to overwrite existing classifications.

Validate and Activate Classification Rules

To validate and activate classification rules:

  1. Create a Classification Rule Set , then add classification rules to the set.

  2. On the rule editor, click Activate.

  3. (Optional) To overwrite classifications, enable Overwrite classifications for <selection>.

    This option lets you overwrite existing classifications for affected keys.

    See Rules Page for a definition of this option.

recommendation-more-help
46b8682c-fda6-4669-9355-1a44923e549e