Entries from June 2020

Rule your Inboxes with RAcS (Read/Action - Spam)

Monday, Monday. You are eager to work on a manuscript. But you have to attend four hours of seminars. And the Email inbox is populated with 127 new messages! If you now start classifying, reading, deleting, and answering Emails, you are not a scientist but a receptionist.

Understanding your Email traffic

Email is currently the most used medium in business communication. It replaced paper letters and fax. You can send large documents quickly and safely. Thus, Email is hard to ignore, and people (ab)use this channel to get your attention or to send you homework. To separate the wheat from the chaff with automated filters, we first need to analyze incoming messages. Initially, I classified them into five categories: 

  • Information - Emails that contain useful information, but do not require immediate action.
  • Spam - Spam includes all types of undesired communication, such as commercials, scam, social network notifications, and extensive mailing list discussions.
  • Homework - Messages that need my attention and response in a reasonable time.
  • Conversation - Exchange of ideas and opinions with one person or a group.
  • Urgent - Emergencies that need my fast decision or intervention.

After careful thinking, I reduced the categories to:

  • Read/Action (=Ham): The Emails you need to read because they contain essential information or require your action.
  • Spam: The Emails that you can safely ignore.

This binary classification facilitates the automated filtering of messages by artificial intelligence (explained below).

Unifying inboxes and Contacts

You probably use more than one Email account and different devices. Unifying inboxes simplifies the organization of your data. Thus, I tested various commercial and free Email programs supporting multiple IMAP accounts and a unified inbox (Evolution, Mailpile, Mailspring, Nextcloud Mail, Rainloop, Thunderbird). Finally, I chose the web client Cypht, which I installed on a virtual personal server (VPS). Therefore, I can use the same interface and information on all internet devices. 

On Android, I made good experiences with K-9 mail. K-9 mail also implements OpenPGP (https://www.openpgp.org) for Email encryption and verification. Of course, nobody cares about privacy.

I store contacts manually using a private LDAP server. This solution is more laborious than using Google contacts and automated address books. However, Cypht provides LDAP integration, and maintaining valuable contact data is worth the effort.

Technical helpers

Warning: This section is for computer nerds. Ordinary people activate the spam filtering of their Email provider and move to the Habits section below.

Advanced spam filtering

Too many undesired Emails still passed my providers' (IONOS, Microsoft Office 365 Outlook, and Gmail) filters. Thus, I needed a more sophisticated tool. The Apache SpamAssassin provides a rich set of efficient spam filtering methods, such as DNS and URI blacklists and Bayesian filtering. 

I installed IMAP Spam Begone (ISBG) on my VPS. ISBG checks the IMAP accounts every 15 minutes and moves messages above the defined score threshold to separate spam folders.

By default, the SpamAssassin efficiently removes real spam, such as phishing scams and commercials from suspicious servers.

However, there is still another source of spam: Internal discussion lists, social network notifications, newsletters, and product information, i.e., messages from reliable senders, but with irrelevant content. You can train the Bayesian filter of SpamAssassin with Spam and Ham messages. For this, I created a separate SpamAssassin Email account with Spam and Ham folders. I move Emails that belong to one of those categories into these folders, and SpamAssassin scans them daily (cronjob 'sa-learn') to improve its classification performance. After some training, most bothering Emails get eliminated.

Archiving and clean-up

You can boost the performance of IMAP accounts by archiving and deleting obsolete Emails, e.g.:  

  • Save Emails older than 180 days locally, and delete them on the IMAP server: Folders InboxSentHam
  • Delete Emails more past than five days: Spam folders

Both tasks can be performed with the program 'archivemail', creating one mbox file for each archived folder. The Email archives are backed-up on a remote server. Note: Probably, this archiving strategy is paranoic. 

Habits

The ISBG/ SpamAssassin eliminates most of the automated mass Emails. 

Now we have to adopt effective habits:

  • Unsubscribe Email newsletters and notifications. Do not fetch RSS feeds with your Email client.
  • Don't worry about urgent Emails you might miss. If something is really vital, people will find you. 
  • Be careful about accepting homework. If a student sends you a Ph.D. thesis on Friday, because (s)he needs the corrections on Monday, you got all the weekend! If possible, delegate work.
  • Avoid Email chatting. "It shouldn't take 30 emails to schedule a 30-minute meeting" (https://doodle.com/).
  • Don't waste your time with sloppy Emails. If the subject seems relevant, ask for complete information.
  • Check your Inbox only once/day. This rule is a difficult one but increases your productivity. You might consider even larger intervals.
  • Write concise Emails. With a descriptive subject, the adequate recipients, and a precise definition of expected feedback.

My Email workflow is:

  1. Move obvious spam messages to the Spam training folder of the SpamAssassin.
  2. Glance through and delete informative Emails.
  3. Read Emails with relevant subjects.
  4. Flag Emails that need your answer or action. 
  5. Answer flagged Emails. Give preference to well-prepared tasks, such as quick yes/no decisions, or documents ready to sign. People will learn.  

Congratulations!

Did you follow this strategy?

Great!

Your incoming Emails reduced significantly (-75%). As well, you drastically shortened the time you spend reading and answering messages.

Now, stay firm and educate people to use Email wisely!


Page 1 of 1, totaling 1 entries