Core Categorizations [2023]

Overview:

What are Categorizations?

• Categorizations are an Alli Data tool that allows teams to create rules that bucket data into reporting dimensions using simple “if, then” logic, replacing the need for complex CASE SQL statements. Learn more here.

What are Core Categorizations?

• Core Categorizations are a set of standardized reporting dimensions & outputs that must be applied across every client. They can be found by navigating to Alli Data > Reports > Categorizations and filtering to those prepended with “core_channel”

How are Core Categorizations used?

• Core Categorizations are automatically pulled into all applicable Data Library tables. Note, they can only be applied to those tables that include the most specific Category Field used by the Categorization.

  • For instance, “core_social_audience_strategy” utilizes the adset_name category field. Therefore, the audience_strategy dimension will only exist in Data Library tables that include that minimum level of granularity.

• In addition to being available for any and all automation & reporting needs, these categorizations are used to create cross agency dashboards & benchmarks.

What are the existing Core Categorization Dimensions & Values?

• The list below reflects all the currently deployed Categorizations (per channel) and allowed output values (see “Output Values” in “Implementation Details” section below for additional information):

• The output values ate also listed here

How can I QA the output of Core Categorizations?

• Follow the link in the Core Categorization description to a Redash report.

• Refresh the query by clicking the bottom left triangle button. By default, the query will pull in all impression driving items from the last year, along with the corresponding Categorization output.

• If a row contains a “N\A” categorization value it will be flagged as "Needs to be Updated". Refer back to the accepted Output values above and update the Categorizations as needed.

How do I update Core Categorizations?

• Work with your Client team leads to identify who is best suited to to update the Categorizations for each channel.

• Navigate to your Client in Alli Data. Click Into Categorizations in the top header. 

• Find the Categorizations labeled as “core_[channel]_[description]” & update for each one for your channels.

• Any published changes made to the default Categorizations will be immediately reflected in the QA dashboard after refreshing.

Implementation Details:

General:

• Categorizations function as IF THEN statements. To update, simply refer to your channel naming conventions and translate those rules to the Alli Data UI.

  • It may be helpful to first navigate to the QA dashboards (see step 3 above) and export the list of active items that need to be Categorized for your channel.

• Leave the Default Value as blank. This will help with initial & ongoing QA of the output.

• Click “Publish” after any updates for the Categorizations to become usable.

Category Fields: 

• The rules can only use one of the pre-selected Category fields listed at the top of the Categorization page (see screenshot below)

• You do not need to define rules for each of the Category fields. Simply use the one most applicable one to satisfy a given dimension.

• When creating the rules, simply click the dropdown to change the field a rule uses. 

• Do not remove any unused Category fields from the top section.

Output Values:

• “Then categorize as” values MUST match to one of the output values outlined below. Not all values must be defined if not applicable to your client, however every row in the QA Dashboard must include one of these values for each Categorization. Note the QA Dashboard will include a “review needed” flag if an output is blank or does not match one of the expected values.

Historical Naming Structure:

• Do not worry about adapting the categorizations for historical naming structures. We are emphasizing the QA of active and go-forward campaigns only.

  • If naming conventions of legacy campaigns are straightforward feel free to still add Categorization rules for them. Otherwise, be sure to use the rules to classify them as “N/A”.

  • What is essential is that there be no uncategorized items in the QA dashboards (which pull in all items from the last 30 days).

• Please reach out to the Client Ops team if you are currently or soon will be undergoing naming convention updates and you are unsure on how to best update the Categorization rules.

Be specific & use the delimiters from the naming structure in your rules.

• As much as possible, include the delimiters in your conditions. For example, campaign contains “US” should be campaign contains “_US_” otherwise you may match a non-US campaign because it includes a string with US (such as “plus_sizing”

Be mindful of the order of your categorizations:

• The rules will run from top to bottom, meaning a value will be set based on the first condition that is met for a given row. With that in mind, we recommended that you add your most specific criteria first to avoid it being mistakenly bucketed by broader rules. For example:

  • A “DSA” campaign (that should be categorized as such) might contain both “_NB_” and “_DSA” in the naming. If the first rule is: IF campaign contains “NB” then “Non Brand” followed by a similar “DSA” rule, DSA campaigns will be categorized as Non Brand because that is the first condition that it met.

  • For extra assurance you can use multiple conditions in your rules to avoid issues like the above. For example, the Non Brand rule above would be set as IF campaign contains “NB” AND does not contain “DSA” then “Non Brand”

Adding a multiple conditions:

• To add a second condition that must be satisfied for a given rule (that is, “AND” functionality), simply click the “+” by the first condition.

Using CSV Bulk Exports:

• Using the “Download Template” feature (by clicking the button at the bottom of the Categorization page) will allow you to build the rules using a CSV file instead of the UI. This is particularly helpful when re-organizing the order of rules. 

• To do so, simply Download Template, add your rules (one per line) and drag and drop the file back over the “Rules” area of the Categorizations UI.

  • The file must be saved as a CSV

  • The DIM_METRIC_NAME must be added exactly as it appears in the Categorization UI. 

  • If writing a rule with multiple Criteria add “and” after the first criteria (in column E). Leave the THEN value blank for that first criteria and assign it to the second criteria. Example below: 

• Lookout for fields that excel converted into formulas (e.g. start with “-”). To address, replace the “=” with a single apostrophe

YouTube & Discovery Campaigns:

• If you are not directly managing some campaigns that pipe into your channels’ QA Dashboard, please collaborate as needed with the appropriate stakeholders and apply the rules. Two notable places this may apply:

  • YouTube campaigns (from Google Ads) fall in the Social Channel

  • Discovery Ads (from Google Ads) can fall into any of the channel tables based on the reported AdNetwork placement.

Misc FAQ:

• Categorizations are not case sensitive.

• You do not need to include quotations in any of the fields.

• You do not need to escape special character functions as you would in a SQL statement.

• More information regarding Alli Data Library (including details on what engine dimensions the pre-selected Category fields” line up to can be found here

• Split By: We recommend only using this functionality for more complex use cases and you should not need to change the default setting (set to “split by”).