Input validation reports¶
The quality of the data sources your brand chooses to make available to Amperity matters when it comes to building unified customer profiles because your brand uses those profiles to activate your customers across a wide variety of downstream use cases. More accurate profiles lead to higher activation rates, better match rates, and increased returns on advertiser spend.
Amperity includes a series of input validation reports that help your brand measure the quality of email addresses, phone numbers, and transactions. Use them to quickly identify data quality issues so that your brand can work to resolve those data quality issues as soon as possible.
Note
Input validation reports are meant to be informative and to provide a way for your brand to explore data and discover when data that is made available to Amperity is missing values or has values that do not match the expected format.
Input validation reports do not have pass or fail thresholds and will not stop automated workflows within Amperity.
Some input validations measure against a single semantic tag, while others use a combination of semantic tags. All input validations are returned as a series of columns that describe the quality of your data as it relates to a specific each report, with a row for each data source.
You do not need to run Stitch or have a working customer 360 database to run input validations. Just publish the domain tables to make them available to the Queries page.
Generate reports¶
Input validation reports may be run from the Sources page. At the top of the page, open the Tools menu, and then click Run validations. The report will start running. The length of time it takes for the report to complete depends on the volume of data in your tenant. The completed report is available from the Input validations report notification that appears in the Recent activity pane after the reports are complete.
The Input validation report dialog box shows all of the individual reports that apply to the current state of your tenant. Each report describes what causes the report to be visible, shows all associated semantic tags, and then returns a series of columns that shows the report status for each data source that was part of the individual report. A link to a query is provided. This query opens in the Queries page and can be used to run queries against domain table data that has been published to the Queries page.
Important
Custom domain table data must be made available to the Queries page. Click the Publish to queries link at the top of the Custom domain table section in the Sources page. This process publishes each of the custom domain tables so that it may be accessed by a query. Allow this process to complete before using input validation queries.
Review reports¶
The list of input validation reports that will be available in your tenant depends on the outcomes of running the validations. Each report is available from the Input validations report that will appear in the Recent activity pane after the reports are complete.
Customer profiles¶
Email addresses and phone numbers are validated for the percentage of missing values, correctly formatted values, and the frequency at which unique values occur. For example, for email addresses:
The percentage of customer profiles with missing email addresses.
The sum of email addresses that do not have a valid format.
The sum of unique email addresses by data source.
The uniqueness of email addresses by data source, compared across all data sources with email addresses.
and for phone numbers:
The percentage of customer profiles with missing phone numbers.
The sum of phone numbers that do not have a valid format.
The uniqueness of phone numbers by data source, compared across all data sources with phone numbers.
The percentage of records with phone numbers that do not contain 10 digits.
Transactions¶
A complete transaction with good underlying data has a combination of values that work together to describe what was purchased, when it was purchased, where it was purchased, and so on. Incomplete transactions are shown using a variety of reports. For example:
All items should be associated with a unique product ID.
All orders should be associated with a store ID.
All orders should have a quantity, a revenue amount, and an order date.
All orders should have a unique order ID.
An item cannot be returned and canceled.
canceled item revenue should be negative
Discount amounts should be positive
Discount percent should be “discount amount / list price”
From which store was a purchase made?
Quantity should be negative when items were canceled or returned
Quantity should be positive
Returned item revenue should be negative
Revenue should be “list price - discount amount”
Revenue should be positive
Through which channel was a purchase made?
Was an item canceled or returned? This should be true or false.
What percentage of product IDs are unique?
Each individual validation report shows the data sources that are associated with the report, the semantic tags that were validated, and the results of the validation. Explore the data that is associated with this report by clicking the Validation query link, which opens in the Queries page.
Privacy rights¶
A privacy rights workflow can help your organization stay in compliance with data protection and privacy regulations, such as those required by California Consumer Privacy Act (CCPA) or General Data Protection Regulation (GDPR).
Privacy rights workflows include the following input validation reports:
Missing request IDs
Missing request types
Missing request strategies
Phone number validation
Request strategy should be connected_pii or exact.
Request type should be delete, delete_pii, or dsar.