Don't Suck At Email(markosullivan.ca)
markosullivan.ca
Don't Suck At Email
http://markosullivan.ca/dont-suck-at-email/
6 comments
Instead of using 'unread' another option is to archive all messages and leave only in the inbox the pending ones.
Or stop using your inbox as a todo list. You wouldn't give people the ability to insert Viagra jokes into your normal todo list; why would you let them send whatever they wanted to a todo list for your most pressing issues?
So, instead, you put "reply to mail xyz" on your todo list for every email you want to respond to but can't respond to right away?
Seems like to many steps to me.
Seems like to many steps to me.
If you can do it in less than five minutes, do it now. If it takes more than five minutes, add it to your todo list (or copy and paste the subject line).
Someone is using GTD :)
As I remember it, GTD does suggest having multiple to do lists based on where you would do each type of task. I think having your emails to do list in your email makes sense. If responding to the email requires you to do a lot away from the computer, than it makes sense to put it on your other to do list.
Also, I'm not sure I'd like to have email subjects, etc... on my to do list. I don't usually keep my list that granular since a lot of the tasks I need to do are pretty amorphous until I actually start on them and figure things out. Maybe its not the best way to do it though....
As I remember it, GTD does suggest having multiple to do lists based on where you would do each type of task. I think having your emails to do list in your email makes sense. If responding to the email requires you to do a lot away from the computer, than it makes sense to put it on your other to do list.
Also, I'm not sure I'd like to have email subjects, etc... on my to do list. I don't usually keep my list that granular since a lot of the tasks I need to do are pretty amorphous until I actually start on them and figure things out. Maybe its not the best way to do it though....
Yeah, I love this method, but if your in a high email environment it requires you to be ruthlessly efficient.
Instead, I flag emails I want to respond to/come back to soon. Also, any emails either a) directly to me (i.e. not to a mailing list I'm on) or b) from my manager(s) gets automatically flagged. Once a day I go through and clean up all my flagged emails.
Instead, I flag emails I want to respond to/come back to soon. Also, any emails either a) directly to me (i.e. not to a mailing list I'm on) or b) from my manager(s) gets automatically flagged. Once a day I go through and clean up all my flagged emails.
I do the same thing. It makes it easy to see, at a glance, what emails still need a response or some action on my part since I sort by date, not status.
It seems that there is more work associated with your option.
If A = all emails, P = number of pending and A>>P, then
his method: click "unread" P times
yours: click "archive" A-P times
A-P >> P
His method is less work.
his method: click "unread" P times
yours: click "archive" A-P times
A-P >> P
His method is less work.
Assuming maximal interface fluency with GMail, you actually perform [on the order of] the same number of input actions for both. Remember, in order to take care of the messages, there has to be a point at which you initially read them. You can then combine the read + file steps into a single triage step.
In triage method A (his method), you first page through all your unread mail (1 click/keystroke to open the first, then A - 1 clicks/keystrokes to advance to the next conversation without returning to the inbox, and 1 final click/keystroke to go back to the inbox.) Then, once you've seen them, you select the mail you want (P clicks requiring dexterity, or A+P keystrokes; it's a compromise) and mark them as read (1 keystroke.) M_a = O(A + P) = O(A). (Because, as you said, A >> P.)
In triage method B (the "naive" method), you page through all your unread mail, and at each step, you either advance to the next conversation, or archive the current conversation and advance to the next conversation. (1 click/keystroke either way.) M_b = O(A).
In triage method A (his method), you first page through all your unread mail (1 click/keystroke to open the first, then A - 1 clicks/keystrokes to advance to the next conversation without returning to the inbox, and 1 final click/keystroke to go back to the inbox.) Then, once you've seen them, you select the mail you want (P clicks requiring dexterity, or A+P keystrokes; it's a compromise) and mark them as read (1 keystroke.) M_a = O(A + P) = O(A). (Because, as you said, A >> P.)
In triage method B (the "naive" method), you page through all your unread mail, and at each step, you either advance to the next conversation, or archive the current conversation and advance to the next conversation. (1 click/keystroke either way.) M_b = O(A).
I also use the unread status for stuff I could go back and respond to, I then archive stuff older than 90 days in my inbox. Having 3 months of email I can search on my BlackBerry is handy when I'm away from a computer.
Rule #8: Don't CC people who don't know each other. Use BCC for that. Nothing bothers me more with email than seeing my email listed next to hundreds others belonging to people I don't know. It's a personal privacy thing. I don't know these people. I don't want them to have my email address.
A minor addition to #2, in which he says that you should only have one question:
If you communicate with someone who only bothers to skim your email (tech support and sales people rarely impress here), or with people with limited reading comprehension for any other reason:
State that one question three times throughout the email, each formulated in a different way but with the exact same meaning.
If you communicate with someone who only bothers to skim your email (tech support and sales people rarely impress here), or with people with limited reading comprehension for any other reason:
State that one question three times throughout the email, each formulated in a different way but with the exact same meaning.
His point #2 is only have 3 sentences in TOTAL, and make the last one the one question you have for them.
Your comment is by no means a minor addition to #2, but more of a refutation.
Your comment is by no means a minor addition to #2, but more of a refutation.
I do not refute it. In the general case, you really should stick to three sentence total and possibly one question at the end.
I provided a special (but annoyingly common) case where it is useful to diverge from the rule. The case is where there is a low probability that the reader will understand any given sentence.
Given that the reader has an estimated 60% chance to understand any given sentence, if you repeat the content in three different sentences, you get a 94% chance that at least one of them will be understood.
As an actual example, I give a shortened transcript from a recent discussion with PayPal support. My first 3 emails were all variations of:
I provided a special (but annoyingly common) case where it is useful to diverge from the rule. The case is where there is a low probability that the reader will understand any given sentence.
Given that the reader has an estimated 60% chance to understand any given sentence, if you repeat the content in three different sentences, you get a 94% chance that at least one of them will be understood.
As an actual example, I give a shortened transcript from a recent discussion with PayPal support. My first 3 emails were all variations of:
I keep getting "The card cannot be used, try again
later" when trying to pay. This has been going on
for months. What should I do?
Hi. Thank you for the answer, but since you also
recommend "just try again at a later time" it seems
you are missing a vital point I'm trying to make:
I have been "trying again" for months.
I have been trying again weeks later several times.
This has been going on for months.
As you recommended, I tried repeating the process at
a later date, but this did not help.
[skip 5 more reformulations]
Please let me know typically how long this "security-
related" issue takes to pass through PayPal's bowels.
How much "later time" do I have to wait to try again?
[signature]
Another thing: don't cc everyone that could possibly want to know about your email's existence. That annoys me to no end.
The email thing that infuriates me is when people send something to a mailing list that I am on, that they think that I should read and copy the mail to me as well. At work, we have people sending the same mail to three mailing lists and then I receive a fourth copy too. Aarrggh!
For a tool that has been around for the last 30-odd years, I am amazed that people are as bad at email as they are. And I'm not talking here about secretaries, or non-technical staff. I'm talking about scientists in a research institute!
For a tool that has been around for the last 30-odd years, I am amazed that people are as bad at email as they are. And I'm not talking here about secretaries, or non-technical staff. I'm talking about scientists in a research institute!
I blame the mail clients. Most don't have a "reply to list" feature, so respondents end up "group reply"ing in order to make sure the mailing list features in the list of To: addresses.
It's lazy, and I hate it, but it gets the job done for them.
It's lazy, and I hate it, but it gets the job done for them.
The clients have degraded because many list managers are inconsistent with how headers on list messages are set.
http://www.unicom.com/pw/reply-to-harmful.html
There isn't any way, for a given client, to ensure consistent behavior when replying to all lists, and for the lists there isn't any reasonable expectation that all users will have the same client behavior. Things are a mess all around now.
http://www.unicom.com/pw/reply-to-harmful.html
There isn't any way, for a given client, to ensure consistent behavior when replying to all lists, and for the lists there isn't any reasonable expectation that all users will have the same client behavior. Things are a mess all around now.
Mutt lets you define mailing lists, so if you press list reply (Ctrl-l by default) it will know whether to reply to the address the message was sent to.
In a work setting I think it's fine as long as there's some level of interest. Obviously it's a balancing act, but I like to keep my direct manager and peers working on the same component informed about test plans, etc.
One other issue with CCing everyone is that you always end up with someone throwing their two cents in about minor issues when they aren't even invested in the project.
One other issue with CCing everyone is that you always end up with someone throwing their two cents in about minor issues when they aren't even invested in the project.
That's why you can have more than one email in the 'To' field:
To: You want to read this.
CC: You may want to refer to this.
To: You want to read this.
CC: You may want to refer to this.
Good luck with that. People at my company regularly send out meeting invites to 3-4x as many people as are actually desired at the meeting. As with email ccs, It's a CYA thing, to prevent other people from claiming ignorance when the shit hits the fan later.
If you have taken the time to send an email, give the recipient time to read it. Do not send the email and immediately walk over or send and IM asking the same question.
I guess the days are gone when top posting is considered bad.