2,433 Unread Emails Is An Opportunity For An Entrepreneur(techcrunch.com)
techcrunch.com
2,433 Unread Emails Is An Opportunity For An Entrepreneur
http://www.techcrunch.com/2008/03/23/a-crisis-in-communication/
19 comments
What if you knew the depth of the mail queue of the person to whom you were writing, before you wrote the e-mail? E.g., you are writing an e-mail to your boss, and when you start writing, you see that he or she has 150 unread e-mails. Might you change your mind and decide that it either 1.) wasn't that important after all, or 2.) important enough to relay the message in person or on the phone?
Not a bad idea!
IDEA: A Karma "bank" for email.
I set up an account at emailkarma.com (sadly, taken, by Xobni, no less!) as a recipient. I set a default free karma amount (say, 100) for new senders, a default karma amount for unsigned emails (say, 50), and route all my email through the service. I also set up a whitelist of known good senders with an infinite karma balance.
People who email me for the first time will automatically be issued 100 karma points by the system. They have one of two options:
1. People who don't know about the karma system can email me directly, but their emails will "spend" the default amount for unsigned emails. When their balance reaches zero, their emails get de-emphasized.
2. People who know about the karma system can choose to spend their remaining balance with me by tagging the email subject with [karma: pointvalue]. Unimportant emails can be assigned less value, and if they absolutely need to email me, then they can spend more of their karma.
Incoming emails are sorted by date and by karma spend. I can reward relevance by giving them more points to spend, and/or by setting their default karma spend to a higher value. I can punish stupidity by adjusting karma accordingly. Webmail interface, or a thunderbird plugin, maybe?
Senders can view their karma balance for a given recipient and default spend (both uneditable) by emailing the service to prove identity and signing up.
Arrington, for instance, can reward good tipsters with more karma, and punish PR trolls by slapping them with a karma penalty.
You could apply this to other forms of communication, too-- your facebook/friendfeed, twitter, etc., though those would have to be passive (relevance weighted by karma balance, no discrete spend available per post/recipient).
I set up an account at emailkarma.com (sadly, taken, by Xobni, no less!) as a recipient. I set a default free karma amount (say, 100) for new senders, a default karma amount for unsigned emails (say, 50), and route all my email through the service. I also set up a whitelist of known good senders with an infinite karma balance.
People who email me for the first time will automatically be issued 100 karma points by the system. They have one of two options:
1. People who don't know about the karma system can email me directly, but their emails will "spend" the default amount for unsigned emails. When their balance reaches zero, their emails get de-emphasized.
2. People who know about the karma system can choose to spend their remaining balance with me by tagging the email subject with [karma: pointvalue]. Unimportant emails can be assigned less value, and if they absolutely need to email me, then they can spend more of their karma.
Incoming emails are sorted by date and by karma spend. I can reward relevance by giving them more points to spend, and/or by setting their default karma spend to a higher value. I can punish stupidity by adjusting karma accordingly. Webmail interface, or a thunderbird plugin, maybe?
Senders can view their karma balance for a given recipient and default spend (both uneditable) by emailing the service to prove identity and signing up.
Arrington, for instance, can reward good tipsters with more karma, and punish PR trolls by slapping them with a karma penalty.
You could apply this to other forms of communication, too-- your facebook/friendfeed, twitter, etc., though those would have to be passive (relevance weighted by karma balance, no discrete spend available per post/recipient).
I agree, and YC is eager to hear proposals from startups working on cures for email overload.
I've detailed how I manage this here: http://www.jgc.org/blog/2008/03/first-assume-all-new-email-i...
But the main problem with the POPFile sorting approach is training. Any automated solution needs to be seamless. People who use POPFile with GMail get automagic training just by moving between IMAP folders/changing labels on messages. I think a plug-in that did automatic email sorting based on some machine learning approach plus some heuristics that was built into Outlook would be compelling.
But the main problem with the POPFile sorting approach is training. Any automated solution needs to be seamless. People who use POPFile with GMail get automagic training just by moving between IMAP folders/changing labels on messages. I think a plug-in that did automatic email sorting based on some machine learning approach plus some heuristics that was built into Outlook would be compelling.
The best tool at dealing with email overload currently is GMail. [edit: removed lengthy explanation of why GMail is a good solution to this problem, if you haven't figured it out you're not fully utilizing GMail ]
I'm sure there is room for improvement and I'd like to see any solutions that come out of YC (but I'd like to see someone actually who uses existing tools designed to tackle this problem explain clearly their shortcomings).
I'm sure there is room for improvement and I'd like to see any solutions that come out of YC (but I'd like to see someone actually who uses existing tools designed to tackle this problem explain clearly their shortcomings).
you just added another 200 applications to the next funding cycle. Hope your inbox can handle it.
I'd love to try to tackle a difficult but noble challenge like this, but I myself do not suffer from email overload (having a simple life where no one cares what you think has its perks) so I'm really not able to empathize with people like Scoble or Mike Arrington. I'd probably do a crappy job of things.
I'm not so sure about this one. There's massive selection bias: people with the ability to complain to thousands of people are exactly those most likely to be suffering from email overload. It certainly seems to be a growing problem, but I haven't heard my mom, dad, or sister complain yet. I'll let y'all know when they do. :-)
A data point in point: I'm working for a very big company at the moment.
The biggest span of control (direct reports) is 12. The number of people with an indirect span of control of greater than 200 people is about 100.
...mind you, that 100 have a lot of money to throw at solutions. Which is why they already have secretaries.
The biggest span of control (direct reports) is 12. The number of people with an indirect span of control of greater than 200 people is about 100.
...mind you, that 100 have a lot of money to throw at solutions. Which is why they already have secretaries.
The thing with email is that it does not require much of a time commitment from even the sender.
Older methods of communication like a phone call or in person meeting require the person who initiates the conversation to invest some time in getting what they want from the conversation.
Let's assume that the majority of the emails received are primarily for the sender's benefit (ex. "please tell me what you think about x for my project" "Could you do an interview for y?"). These messages should get the lowest priority and could even be deleted if they can't be answered within a week or two.
The next set of emails are of mutual benefit to the sender and receiver (such as conversations between clients and companies, friends to friends, and productive internal corporate communication).
The final category is emails for the receiver's sole benefit. Automatically generated emails should be in this category, if they are to inform you that your credit card may have been compromised. Most emails here don't require a response from the receiver. Spam and marketing materials do not fall into this category.
Of course in each category you could drill down by priority depending on the situation.
Older methods of communication like a phone call or in person meeting require the person who initiates the conversation to invest some time in getting what they want from the conversation.
Let's assume that the majority of the emails received are primarily for the sender's benefit (ex. "please tell me what you think about x for my project" "Could you do an interview for y?"). These messages should get the lowest priority and could even be deleted if they can't be answered within a week or two.
The next set of emails are of mutual benefit to the sender and receiver (such as conversations between clients and companies, friends to friends, and productive internal corporate communication).
The final category is emails for the receiver's sole benefit. Automatically generated emails should be in this category, if they are to inform you that your credit card may have been compromised. Most emails here don't require a response from the receiver. Spam and marketing materials do not fall into this category.
Of course in each category you could drill down by priority depending on the situation.
If we had a good way to send and receive money through the Internet then it might make sense to charge a fee to accept an incoming email. One would have a whitelist of senders who get automatic refunds. Everyone else will be spending real money to send you email.
A tiny "postage" fee might also solve spam.
A tiny "postage" fee might also solve spam.
I think that you might be on to something with the fee in the sense that you could apply a solicitation surcharge to unapproved senders. It would only work if it was guaranteed to get your response, otherwise it's not worth it.
What about a separate email address (or even domain) for premium email? You could set the fee to receive an email (and adjust it to restrict flow to what you can handle), and refund people who don't get a response from you within x period of time. Certain senders could be excluded from paying. The senders could even choose how much to pay (These emails were sent without payment. Those were.)
You then publish only that address on your site/blog/whatever, and give your personal address to friends and family.
This idea needs a lot of digging into to really see if it's viable but I like the concept. The potential downside is that the government might like the idea of email tax.
What about a separate email address (or even domain) for premium email? You could set the fee to receive an email (and adjust it to restrict flow to what you can handle), and refund people who don't get a response from you within x period of time. Certain senders could be excluded from paying. The senders could even choose how much to pay (These emails were sent without payment. Those were.)
You then publish only that address on your site/blog/whatever, and give your personal address to friends and family.
This idea needs a lot of digging into to really see if it's viable but I like the concept. The potential downside is that the government might like the idea of email tax.
anshu jain, who runs Deutsche Bank's investment banking arm had 2 full time PAs who read his email and filtered accordingly. They were paid 7 figures and were very smart. The top business people, I presume, don't read their email directly. It's all filtered.
As for Mike A, I presume a huge inbox is party because he's running a huge and growing business, and has leveraged himself to the max. In that sense, it's no bad thing.
As for Mike A, I presume a huge inbox is party because he's running a huge and growing business, and has leveraged himself to the max. In that sense, it's no bad thing.
The unmet demand may be closer to 10,000 a day for "high volume communication" and knowledge workers because there is a "shadow backlog" caused by the poor performance of the current approach. There are a couple of different categories of interaction:
I don't want to lose a key e-mail from someone new (e.g. a prospect) or someone I want to reconnect with (e.g. an old friend) in the midst of everything else
I want to have some answers suggested based on a collection of blog posts, a FAQ, or some other knowledge repository.
I want to recombine a number of e-mails into a complex branching thread because that's a work flow that a team I am a member of embraces (see http://www.43folders.com/2008/03/12/patterns-email-conversat... ) so that I have the full context for the conversation.
I want to make sure I don't lose touch with folks I have had a prior shared success with (I would like to boost my effective dunbar number).
There are several others. Please feel free to contact me directly if you would like to continue the conversation. I think this is one area that is overdue for a significant change in paradigm, the evidence that the current approach isn't scaling can be found in a variety of secondary and tertiary coping behaviors that we have come to accept without attacking the root cause. I think there will be many approaches that continue to use SMTP as a transport protocol but enhance the user interface with embedded analytics and automation, as well as adding alternative forms of response to include blogging, wikis, IM, and VoIP.
see also http://news.ycombinator.com/item?id=143878
I don't want to lose a key e-mail from someone new (e.g. a prospect) or someone I want to reconnect with (e.g. an old friend) in the midst of everything else
I want to have some answers suggested based on a collection of blog posts, a FAQ, or some other knowledge repository.
I want to recombine a number of e-mails into a complex branching thread because that's a work flow that a team I am a member of embraces (see http://www.43folders.com/2008/03/12/patterns-email-conversat... ) so that I have the full context for the conversation.
I want to make sure I don't lose touch with folks I have had a prior shared success with (I would like to boost my effective dunbar number).
There are several others. Please feel free to contact me directly if you would like to continue the conversation. I think this is one area that is overdue for a significant change in paradigm, the evidence that the current approach isn't scaling can be found in a variety of secondary and tertiary coping behaviors that we have come to accept without attacking the root cause. I think there will be many approaches that continue to use SMTP as a transport protocol but enhance the user interface with embedded analytics and automation, as well as adding alternative forms of response to include blogging, wikis, IM, and VoIP.
see also http://news.ycombinator.com/item?id=143878
If you read that and think "I have 24 emails a day; he must be getting about 100x as many pings for his attention," it might be worth reconsidering how people react to his email policy. If I wanted to get in touch with this guy, I wouldn't send an email. I'd send five, spaced an hour or two apart, with varying subject lines. One joke, one brief and ambiguous ("re:requested") one with caps, etc.
He'd probably be getting 400 emails a day if he didn't force people to resend anything they expected him to read.
He'd probably be getting 400 emails a day if he didn't force people to resend anything they expected him to read.
Or an opportunity for a secretary?
Even with a secretary it would be hard to manage 1,000 e-mails per day, much less 10,000 using the current e-mail clients. This would also be true of outsourcing it to a team of hardworking Estonians, Irish, Indians, or Canadians. Look at electronic discovery techniques (reading e-mail in a lawsuit evidence discovery process) where you may pay dozens of attorneys to sort through e-mail as another limit case of throwing bodies at the problem.
Would an urgency metric be a solution, allowing you to sort those 2433 emails to see the 50 that you really should reply to?
That would require you to sort each email based on urgency, and the only way to do that (effectively) is to open and read the email. That's the problem.
Point taken. Perhaps some combination of the sender and subject could be a reasonable approximation? In my own experience, the sender almost exclusively determines the importance or urgency of any email.
That's a compelling idea.
The simple solution: A mail client that doesn't say how many unread messages you have. If it's urgent they'll find another way of contacting you.
This is one of the great things about Twitter. Nowhere is there a "x twitters unread" on the site. You simply dip in and out as you wish.
This is one of the great things about Twitter. Nowhere is there a "x twitters unread" on the site. You simply dip in and out as you wish.
Hey kids, how about a contest?
Someone (not me) could create a control data base of say 2,433 emails. The data base could be downloadable.
Whoever feels like it could work on their solution and in x weeks or so post either a url for their new app, a proposal, or even a report based on the data. The rest of us could vote just as we normally do and the top 3 contestants could compete for the top prize of ________.
(I may be onto something, but my creativity is waning. A little help please. Any ideas?)
Someone (not me) could create a control data base of say 2,433 emails. The data base could be downloadable.
Whoever feels like it could work on their solution and in x weeks or so post either a url for their new app, a proposal, or even a report based on the data. The rest of us could vote just as we normally do and the top 3 contestants could compete for the top prize of ________.
(I may be onto something, but my creativity is waning. A little help please. Any ideas?)
i really like that idea actually. having a controlled and constant data set, could provide a lot of insight.
If any system (attempting to solve this problem) allows one to add 2,433 emails to an "inbox", then it is doomed to fail.
Sometimes, it's the tool user, not the tool itself, that is to blame.
Sometimes, it's the tool user, not the tool itself, that is to blame.
In the world of vitamin problems, this is definately a pain killer. I wish I had time to work on a second startup and tackle this as well. Best of luck to those who do - I'm eagerly waiting.
How about a Bayesian solution that observes how you answer emails, and prefills a reply (or automatically puts emails into trash) based on your historical reaction to similar emails?
Dear Mom,
Thank you for your recipe for fairy cakes. Thanks to my previous responses to e-mails containing the word fairy here are some links to soft core homosexual porn.
Yours,
Soon to be out of the closet Joe, who no longer trusts the guy tuning his auto-mail algorithm.
:)
Thank you for your recipe for fairy cakes. Thanks to my previous responses to e-mails containing the word fairy here are some links to soft core homosexual porn.
Yours,
Soon to be out of the closet Joe, who no longer trusts the guy tuning his auto-mail algorithm.
:)
Hire an offshore personal assistant to read/sort your email?
I'd love to see something along the lines of http://cr.yp.to/im2000.html
I think this problem is less of AI, more of HI; not so much a computing problem but more of a cognitive one.