Percentages Of E-mail Clients By OS And By Feature?
Krotus asks: "I've been looking for information on user shares of popular e-mail clients, and am really surprised at how hard it is to find anything beyond vague speculations. I suspect that lots of us could use these numbers, but has anyone been able to find out about what percentages of users are using Eudora, Outlook's various versions, mutt, pine, mh, Notes, etc.? Almost as useful would be numbers on what percentage of clients can parse HTML, vs. plaintext or something proprietary.
Forrester, Jupiter, and IDC have all come up blank. The question has been asked at Abuzz with no luck. Maybe the collective knowledge of Slashdot will be more fruitful?"
Despite the fact that the review process was very flawed and a total con (typical of consultants), it very clearly outlined that there is a huge mix of both email clients and email servers out there and that we really should look at standardising the clients and consolidating the servers.
Part of the process was a survey of current usage across campus. The client side results were (from memory, in no order and staff only):
The students mostly use Eudora, although the use of our internal student webmail system is starting to increase quite dramatically.
There is also quite extensive use of free web based email (such as hotmail), but I don't have any stats on this. We are actively discouraging its use, but the Uni doesn't have any real balls when it comes to making policy about it.
The really sad bit about all this is that the review process somehow decided that Exchange is the best server solution. Can anyone say Linux box running SMTP/POP/IMAP?
Why are you surprised that it's hard to find any numbers? It's not like webservers, or even web browsers, where you can query the server or look at log files. Emails pretty much go straight from user-to-user. AFAIK, mail servers don't bother logging X-Mailer headers. Really, the only way to find out what clients are in use is to actually ask people in a survey, which probably wouldn't cover too many users.
hey! the problem is unlike the web it's hard to gather stats on email clients, right? So, let's just like send this email around asking people to put their email client on the list at the top, and forward it on, and every hundredth can cc it to a central location....
;)
...waitaminute.
Seriously, tho, you could get rough numbers on Outlook users from the LoveBug stories I'd wager.
Returned Peace Corps IT Volunteer
I personally use BeOS's own BeMail and it's database-like directory struc for everything important to me...though there are other clients availabe (some Outlook-like), I much prefer the simplistic approach.
"AOL, CIA, NSA, whatever, they all collect information, and they are all out to screw the american public"
Well at my place of business we use Eudora, but due to bug problems we may be moving to Netscape Mail. In alot of corporate Environments NEtscape Mail is still popular.
Many people i deal with end up using Outlook Express because it comes with Windows, at least form a windows persepctive. It's fine for base needs.
this space for rent
"Patience is a virtue, afforded those with nothing better to do." - I don't remember
Why not write a cross-platform network worm that spreads via e-mail? Every time it hits a client, it tells a central server what program it finds. It could spread automatically if it finds an Outlook client, and spread on the honor system for anything else. Just ask people to run this EXE file or shell script or visit this web site and see what you get. ;-)
Daniel J. Peng
I don't understand why people don't configure an email server with logging to figure this out -- like so:
1. Find population that you're interested (say mit.edu addresses)
2. have the SMTP damon on mit.edu save the X-mailer headers, devoid of any personal data.
3. Make statistics...
I mean, granted lots of people may use different smtp servers (you'll miss some of the CS folk) but at least it'd be realatively accurate...
Why doesn't forester or some other research group partner with @home, pacbell, and some of the other bigg'o providers and log for a couple of days...
comments?
willis
there is no thing
what else could you want?
I was just thinking -- Eudora, et al are probably really popular, but doesn't AOL have something like 23 million users? (is that users? or accounts). Hotmail and Yahoo also have some crazy numbers as well -- not to mention a lot of other similar webmails (usa.net?) that are probably a bit smaller.
I'm sure in business outlook is pretty big -- but most of the students I know use Eudora or plain'ol webmail... (I myself use pine).
yeah...
willis.
there is no thing
what else could you want?
check it out, yo.
/bin/usr/mail, and if I do, I'm just a player then and I don't need to be counted....)
/usr/bin/mail -- and that might be a fingerprint as well...
The reason why you'd want to check the X-Mailer is to see who is sending the message... but the overall goal is figuring out who READS the message with WHAT client. Therefore, things like your scripts, etc. should be tossed out from the get-go -- they give no useful information about what a particular person is using to READ the message.
(i.e. if I send mail from hotmail, then I probably read it from there too... I probably don't read mail from
>I have used several e-mail clients, and NONE of them have ever put their name in the headers.
grep -h "X-Mailer" ~/Mail/* | sort -u >foo
X-Mailer: Web Mail 3.5.1.4
X-Mailer: Allaire ColdFusion Application Server
X-Mailer: Juno 1.49
X-Mailer: Kana Customer Messaging System 3.0
X-Mailer::1.0 (http://www.gossamer-threads.com/scripts/)
X-Mailer: AOL NetMail version 2.0
X-Mailer: Becky! ver 1.25.07
X-Mailer: AOL 4.0 for Windows sub 30
X-Mailer: eGroups Message Poster
X-Mailer: USANET web-mailer (M3.0.0.70)
X-Mailer: Microsoft Outlook Express 5.00.2314.1300
X-Mailer: QUALCOMM Windows Eudora Pro Version 4.2.0.58
X-Mailer: Mozilla 4.7 (Macintosh; U; PPC)
X-Mailer: Microsoft Outlook CWS, Build 9.0.2416 (9.0.2910.0)
X-Mailer: ELM [version 2.4ME+ PL34 (25)]
X-Mailer: Internet Mail Service (5.5.2650.21)
X-Mailer: MIME-tools 4.104 (Entity 4.117)
there are a bunch more, too. Notice how some of these don't represent a mail client that somebody's actually using (a coldfusion client? sounds like a form to me) and thus shouldn't be counted...
Also, some things clients might need to be teased out (pine, hotmail, aol (v5?), etc.) but they probably all have unique fingerprints like below...
Received: from 12.73.227.x by www.hotmail.com with HTTP
Received: from imo-d09.mx.aol.com
Message-ID:
the only thing that I didn't get a fingerprint on was
(I don't think I know anybody who uses Pegasus -- that might be why there's no peg. client listed. I'm sure there's a fingerprint lurking in its headers as well, same with MUTT, etc.)
willis
there is no thing
what else could you want?
I am looking for the same info. I could only find mid 1998 data (besides the Slashdot poll showing 25% pine usage.)
Client Server Survival Guide, 3rd Edition. Robert Orfali, Dan Harkey, Jeri Edwards. page 408
Derived from data from IDC, Dataquest and EMMS
Other- 19.5% (29.9 M)
Notes 14.3% (21.9 M)
Eudora 11.8% (18.0 M)
cc:Mail 9.8% (15.0 M)
Exchange 9.8% (15.0 M)
Groupwise 8.1% (12.4 M)
Hotmail 6.6% (10.0 M)
Software.com 6.6% (10.0 M)
MS Mail 3.9% ( 6.0 M)
HP OpenMail 3.5% ( 5.4 M)
Suitespot 3.4% ( 5.2 M)
FirstClass 2.7% ( 4.2 M)
Are the biggest and berst in this area. I know ferris had some data out last year, but it didnt go into each of the imap/pop clients. Plus his data came from the vendors who consider Purchased to mean used. Thats why Notes shows over 25Million seats even though I can only count about 6Million thru my clients. I would be very interested in helping, some of the corporate biggies might give me access to numbers faster then they would to people they don't know, I deal with about 20% of the lan based messaging world. -D_ale> www.jconsult.com
I've continued looking in the week since I posted this question, and finally managed to find a place with good extensive data, albeit rather old (October 1998). They've apparently been doing these surveys for years at Georgia Tech. See GVU's WWW User Surveys for the whole thing, and specifically the Technology Demographics section for direct information on my question.