Message deliverability monitoring

A non-transitory computer-readable medium containing instructions, comprising:

• instructions for receiving an original email comprising content code in a markup language;
• instructions for parsing the content code of the original email into a tree;
• and
• until a stop condition is reached:
• instructions for generating a plurality of test emails, wherein the generating of each test email comprises:
• instructions for selecting a subtree of the tree different from other selected subtrees, and
• instructions for converting the subtree into content code in the markup language to create content for the test email;
• instructions for transmitting the plurality of test emails to an email address;
• instructions for monitoring a plurality of sensing signals from the test emails, wherein the sensing signals for each test email comprise whether the test email was opened, whether a link in the test email was clicked, whether an unsubscribe request was received, and the time the test email was received;
• instructions for determining whether each of the plurality of test emails is categorized as spam by the email address based at least in part on the plurality of sensing signals, the determining comprising predicting a change in the categorization of emails as spam by an Internet Service Provider (ISP) providing the email address by determining that the email address does not categorize the plurality of test emails as spam at the same rate as the known rate of spam categorization within a threshold, and
• after determining whether each of the plurality of test emails is categorized as spam by the email address, instructions for determining whether a stop condition is reached by identifying one or more sentences in the plurality of test emails causing the original email to be classified as spam.

A system comprising one or more processors and a memory storing instructions that, when executed by the one or more processors, cause the system to perform the following operations:

• receiving an original email comprising content code in a markup language;
• parsing the content code of the original email into a tree;
• and
• until a stop condition is reached:
• generating a plurality of test emails, wherein the generating of each test email comprises:
• selecting a subtree of the tree different from other selected subtrees, and
• converting the subtree into content code in the markup language to create content for the test email;
• transmitting the plurality of test emails to an email address;
• monitoring a plurality of sensing signals from the test emails, wherein the sensing signals for each test email comprise whether the test email was opened, whether a link in the test email was clicked, whether an unsubscribe request was received, and the time the test email was received;
• determining whether each of the plurality of test emails is categorized as spam by the email address based at least in part on the plurality of sensing signals, the determining comprising predicting a change in the categorization of emails as spam by an Internet Service Provider (ISP) providing the email address by determining that the email address does not categorize the plurality of test emails as spam at the same rate as the known rate of spam categorization within a threshold, and
• after determining whether each of the plurality of test emails is categorized as spam by the email address, determining whether a stop condition is reached by identifying one or more sentences in the plurality of test emails causing the original email to be classified as spam.