Why is Hotmail not as popular as Google?
Why is
Hotmail not as popular as Google?
Hotmail was the foremost free email service when
Microsoft bought the company in 1997.
Microsoft bought them as a market share play for the 9
million user accounts they had at the time — for approximately $400M, or about
$40 per user.
This was back when the race was all about acquiring
users, and what you would do with those users once you had them, in order to
monetize your investment, was “a problem for later”.
In other words: prior to the dot bomb.
One of the major issues leading to the demise of Hotmail
was that Microsoft spent a great deal of effort spinning its wheels.
It did this in order to first put HTTP proxy servers in
front of the FreeBSD servers upon which Hotmail was originally built so it
would look like they were all running on Microsoft technology, and then
replacing the FreeBSD servers on the back end so that they actually were
running on Microsoft technology.
Rumor has it — and I knew casually the people who started
Hotmail through the FreeBSD community, so it’s not that much of a rumor — that
it took nearly 8 times the number of Microsoft servers to replace the FreeBSD
servers.
This was actually money well spent by Microsoft, since it
was Windows first foray into building “web scale servers”.
But in doing this, they lost critical first mover
advantage to GMail.
At the time they lost the advantage, the HTTP 1.1
specification was being finalized.
One of the major features of HTTP 1.1 over HTTP 1.0 was
that there was a field in the request header which included the intended target
of the request.
This may not seem like a big deal, but it was a very big
deal for many companies, including Paul Vixie’s company at the time, which was
trying to provide virtual hosting services using NetBSD on DEC Alpha hardware.
Microsoft, on the other hand, was attempting to leverage
its near-monopoly on the desktop OS to crown IE as the main browser in use on
the Internet, as opposed to Netscape.
And one of the things that Microsoft did to try to
accomplish this was introduce their own “de facto standards” as modifications
to the HTTP protocol, and the HTML content markup language on top of the HTTP
protocol.
Just as when Microsoft initially “missed the boat” on
TCP/IP, when they and Novell were vying for a contract with AT&T to provide
a network backbone, either based on Novell’s SPX/IPX, or Microsoft’s
NetBEUI/NetBIOS — they missed the boat again.
Microsoft was late to the game providing HTTP servers
which conformed to the HTTP 1.1 specification — and therefore capable of
virtully hosting many domains on the same server.
Google did not miss the boat.
As a consequence, Microsoft lost out on feature parity;
GMail was able to host your domain name on their mail infrastructure; Hotmail,
however… was not.
I know this because I was involved in an ambitious
project at another company, which was purchased by IBM, at which point it
became an IBM project.
Hosted Domain eMail.
You get your own domain name… and then you use IBM’s
WebMail servers to access the email on your domain — which you could do,
because of HTTP 1.1.
And that was the “game over” moment for Hotmail.
Over the intervening years, Hotmail continued to lose
ground on feature parity with GMail.
Hotmail was the ugly red-headed step-child that Microsoft
didn’t want to put any effort into improving.
Today, their hosted domain support is inferior, their web
interface is inferior…
…but most importantly, their SPAM filtering, and their
malware filter is inferior.
They also have really awful size restrictions on
mailboxes, inferior large file send support, and they are pretty tolerant of
SPAM being sent from their service (they are major bad actors, in other words).
While things have (somewhat) improved with their
migration to their Outlook web mail — Hotmail itself has been technically dead
for years, now — the Outlook “pod person” which replaced it still has yet to
catch up in many areas.
And Microsoft does not look to be throwing money into any
of these areas.
So blame Microsoft.
Blame:
- Getting caught in the dot bomb
- Trying to own the Internet’s basic networking standards
- Trying to own HTTP
- Trying to own HTML
- Being slow off the mark supporting HTTP 1.1
- Not having server software capable of virtual hosting when browsers started supporting it on the front end
- Etc.
Blame what you want.
It was a cluster …er …of problems.
Comments
Post a Comment