A number of our users have asked for additional information on the "behind the scenes" functionality of ClearContext. This is part 1 in a series of posts summarizing ClearContext design concepts. For more detailed information, see our whitepaper Designing a More Effective Inbox.
ClearContext automatically assigns importance to email based on message characteristics such as the sender of the email, user involvement in a thread, the priority assigned to the email, etc.. Automated settings can be overridden to designate certain senders or messages as higher priority. The result is a prioritized, color-coded inbox with the most important email appearing in red at the top of the message list, enabling the user to quickly view and address important messages.
To perform this prioritization, ClearContext goes through several processes to determine which email message is likely to be most important to the user, including an analysis of contact and email history and assignment of message and thread priority. In this post, we present our Contact Analysis algorithm.
On startup, ClearContext’s patent pending contact analysis algorithm analyzes email history and email contact information (within the Outlook data store) to determine from which email addresses the user sends and receives the most email. Email addresses are then prioritized according to volume, giving the user an accurate picture of his/her most important email contacts.
Contact priority is assessed in the following manner:
1. All email addresses contained within Outlook’s address book are extracted for analysis.
2. Email history is scanned to identify and extract email addresses that are part of the Outlook data store, but are not present in the user’s Contacts.
3. A score is given to individual email addresses based on the number of emails received from that contact, taking into account the age of received email to ensure that more recent email activity is given a higher score than earlier activity.
4. Additional weight is given to these addresses based on the number of emails that have been sent to that contact, taking into account the age of sent mail.
5. A final composite score is calculated using the weighted score components.
ClearContext creates a prioritized list of contacts, placing those email addresses with higher volume at a higher rank. The top percentile ranges of email contacts are placed into three categories, Very High, High, and Normal. These values are assigned within the user’s address book. If an address in the top percentile ranges does not exist in the user’s contacts, an entry is created in a special ClearContext Contacts folder within the Outlook data store.
Once this analysis is complete, the application uses this data, along with additional information about the message, to increase or decrease the priority of an email (prioritization details in an upcoming post). The user can manually change contact priority within Contact records, further tweaking the automated analysis.
Watch this blog for future posts on Inbox Manager design. Next up: Message Prioritization