From: Adam Rifkin (adam@KnowNow.com)
Date: Sat Nov 25 2000 - 23:57:39 PST
Fascinating reading five and a half years later...
Chief Yahoos: David Filo and Jerry Yang
by Mark Holt & Marc Sacoolas
Yahoo (Yet Another Hierarchical Officious Oracle) is the most popular
hierarchical index of the World Wide Web (WWW). Created in 1994 by David
Filo and Jerry Yang, both Stanford University Electrical Engineering
Ph.D. candidates currently on sabbatical. This popular index receives
more than 2 million accesses a day representing over 200,000 users.
This interview was done over two meetings in Palo Alto, California, USA
in May 1995. The first meeting was with David Filo at Caffe Verona and
the second meeting was with David and Jerry Yang at the Blue Chalk Cafe.
BACKGROUND ON DAVID AND JERRY
Q: Are you from the San Francisco Bay Area?
A: (David) I'm from Louisiana. I went to Tulane University in New
Orleans as an undergrad and majored in Computer Engineering. Then I came
out to California to get my Masters at Stanford in EE [Electrical
Jerry grew up in San Jose and went to Stanford for his bachelors. We're
both in Stanford's Electrical Engineering program doing electronic
computer-aided design research.
Q: So is that where you guys met?
A: Yes, we're in the same research group.
Q: Have you known each other long?
A: For about 4 or 5 years.
Q: How long have you been using the Web?
A: Pretty much ever since [NCSA] Mosaic came out in mid to late '93. And
before Mosaic, we used gopher a bit... and of course we used e-mail and
newsgroups before that.
BACKGROUND ON YAHOO
Q: When did you start Yahoo, and did it start out just for personal use?
A: We started Yahoo in about April '94. It started out as a way for us
to keep track of things that we were interested in. We began with a
Mosaic hotlist, and we just added things to it. When our hotlist was
about a hundred items or so, all you could do was scroll through it
sequentially. Soon we couldn't find anything anymore because the list
was so big and there was no way to search it.
So, we created some tools that allowed us to hierarchically categorize
our links much the same way that many browsers like Mosaic and Netscape
do today. Once we did that, finding previously visited links and adding
new links became much easier. We then made the list available via the
Web to make it easier to browse and added the search capability to make
it easier to find specific entries.
Q: Was Yahoo available to other people from the start, or was it just
between you two?
A: It was just between us at first. The tools that we developed
consisted of a few Tcl/Tk and Perl scripts that only we had access to
and knew how to use. After we made it available via the Web interface,
others had access to it but still not many people were aware of it.
Q: When and how did other people find out about Yahoo and start to use it?
A: Well, we never really listed Yahoo anywhere. It was never listed with
any of the "What's New" pages or posted to any newsgroups. A couple of
friends started to use it. It was somewhat slow to catch on initially
because we never really advertised it. It just grew by word of mouth and
by people linking to it from their home pages.
Q: From the start were people able to add links, or were they just
watching you add your stuff?
A: Well, first of all we had very few users. We might have had a hundred
accesses a day. So there was really no demand from the users to add
their own links. Things changed over time though as our access rates
doubled every month. Through word of mouth on the Net more and more
people began using it.
Q: So you never listed Yahoo on any of the other search web sites?
A: No. Although, after a while we automatically became listed with some
of the search engines when their spiders and robots found our site. We
also would refer our site to others who were looking for particular
types of information on the Web. Other users would do this too in
newsgroups and such.
Q: We know how that goes. We've told lots of people about Yahoo. We
found out about it through Web surfing on other corporations'
pages. Probably others found out about it in similar ways.
A: It just kind of slowly evolved. Or maybe not so slowly.
Q: And when did you guys realize that Yahoo was more than just your
personal hotlist? Was that after a couple of months?
A: Probably about six months after we started. Things really changed in
this respect when we added the ability for people to add their own
links. At that point it was obviously no longer just for our own
interests. We also began to add more functionality to make it more
useful to a wider range of users.
Q: Such as searching and graphics?
A: Searching was available from the very beginning, but we did improve
things over time. The main improvements were in its ease of use and its
efficiency. They were necessary improvements as the number of users grew
and the demographics broadened. The graphics were also tweaked over time
to make the pages as efficient as possible in terms of load time.
Q: Are there about 39,000 links right now?
Q: And is it still about 2 million accesses a day?
A: Right. Now it's about 2 million. We used to actually get more that 2
million, but that was with graphics. Then we took graphics completely
out of Yahoo because we only had one DEC Alpha for a server and it was
beginning to peak out in the number of requests it could handle. We
didn't have any extra hardware, so instead of getting more equipment, we
just started taking the graphics out.
Q: So there's probably now about 200,000 to 250,000 users...
A: Yes, actually that's something I've been meaning to check lately. Our
logs are so big that our analysis scripts require too much in the way of
machine resources to run on a nightly basis.
Q: So the reason the graphics were removed was to lower the number of
accesses to your server, not to make Yahoo quicker for user with low
A: Right. Even when we did have images, we kept them very small -- around
a couple of hundred bytes.
Q: What about the size of the menu bar?
A: Even the menu bar was only about 1K. If you have a 14.4k connection,
anything over 1K starts to become noticeable. We didn't think that
graphics were that big a deal for users. When we took the graphics off,
we did get quite a few requests saying, "Leave em' off; it's faster that
way". We also got a few saying, "Put em' back", or "Leave them on
there." Really, our graphics weren't much to begin with.
Q: You mean they weren't Adobe Photoshop quality?
Q: So, Yahoo's growth rate was doubling every month?
A: Well, that was true for most of '94 until the end of the year with
the holidays, which kind of knocked the number of accesses down a
bit. We've continued to grow steadily during '95 but at a slower
rate. Last year most people didn't know about us so we continued getting
lots of new users over time. Now that has changed somewhat and the
number of new users is more in line with the overall growth of the
Web. Even that growth by itself is quite dramatic, especially when the
large online services introduce Web access to their users like Prodigy
and CompuServe have.
Q: Did you guys notice when Prodigy started offering access to the Web?
A: A little bit, but not as much as you might expect. It took awhile for
their users to get onto the Web. In fact, many of their users still
don't use the Web yet. They don't all download the latest software at
the same time when it becomes available. So it takes days or weeks, and
even then you don't see all the activity because of the caching done by
their proxy servers.
Back when they first launched their Web access, we were still tracking
the hosts that accessed Yahoo on a daily basis. On either the first or
second day they became the most popular site to visit Yahoo. That
Q: Are most of your users in the United States?
A: Yes, well certainly the big ones are. It used to be just the
companies like Sun, DEC, HP, and AT&T. And now Prodigy and Netcom bring
the most users.
Q: Do you have more individual or corporate users?
A: Well, that's hard to tell. In reality, big corporations are all
behind a firewall where their users are all routed through one host. So
it is hard to tell how many users there are in corporations.
Q: For instance, Sun has one machine at their firewall?
YAHOO TECHNICAL QUESTIONS
Q: Do you intend to keep Yahoo text based, or, after you set up your
business, will we start to see the graphics put back in?
A: Well, I think graphics will come back. Right now, we're equipment
bound. We have another machine on the way which we could add graphics
with. We will probably also keep the text version for the people who
Q: Would users be able to choose whichever version they wanted?
A: Yes, but I'm not really sure the best way to do this. It's kind of
annoying having to always select a graphics or text version when you go
to someone's home page. It would be nice if you could automatically and
transparently send your preference with each request.
Q: Yes, you could identify their browser when they access your page.
A: That's true, but it still doesn't solve all the problems.
Q: What is the bottleneck for Yahoo? CPU, hard disk, net bandwidth, or
A: We had a 10 Megabit connection to the Net at Stanford, which was
connected directly to the Internet. So bandwidth was never a limit. Disk
space was never a problem either. At Stanford, it ended up being that
the machine could not handle enough requests. Limitations on the number
of TCP/IP connections a single machine could handle was our bottleneck.
Q: Whose HTTP server were you using?
A: We wrote our own. We were using the NCSA server in the beginning,
and, around 100,000 to 200,000 hits a day, it was pretty much just
killing the machine. So, we wrote a server that is extremely barebones.
Q: And that's your own proprietary server?
A: Yes, for now. But, we'd rather use someone else's, like Netsite. We
just haven't had time to switch over.
Q: So your server is really your own customized http server?
A: Right. There's not a lot of incompatibility, but, the cgi stuff is
slightly different. It's the way things are passed that's different. And
there are smaller issues like case sensitivity. Most servers are case
sensitive while our's is not. So little things like that prevent us from
switching right now. With the NCSA server the bottleneck was that it was
a CPU hog. After we wrote our own server, it ended up being a TCP/IP
limitation. Basically, current Unix implementations of TCP/IP can only
handle about 100 connections per second.
Q: So your bottleneck was TCP/IP?
A: Right. What seems to happen is that there are some tables that grow
with the number of connections. Many table lookups are performed using
inefficient linear searches, instead of hashing or something more
intelligent. Once the tables reach a certain size, the machine spends
all its time searching these lists. The fundamental problem is that the
machines were never really designed to handle this type of network traffic.
Q: Web servers weren't around then.
A: Nope. In addition to the TCP/IP problems mentioned, there are
limitations on socket queue size in most of the Unix systems shipped
today. These limits are hit with even a small volume of Web
traffic. Luckily the problem can be solved by patching the kernel to
increase these queue sizes.
Q: So did you rebuild your kernel to make those improvements?
A: Yes, we increased the queue limits as much as necessary to avoid that
Q: Did you create all the search software?
A: We made the search software.
Q: Are your Web server search software and kernel independent, or do
they work together in any ways? Does your http server talk to your
A: Well, yes.
Q: In a way that no other search engine and server that you know of does?
A: Yes. In fact, there is an HTTP server for the non-search stuff, and
there is a separate server that is actually modified for the search
engine. That server has some load balancing capability built in. The
search and server are tightly integrated. The kernel is again just a
patched version of what the vendors ship.
Q: So, over at Netscape, is TCP/IP still the bottleneck?
A: Yes. What we're doing at Netscape is trying to get more machines
running at 100 percent. Right now we are running on 3 machines and we
have a 4th coming in. Netscape is loaning them to us.
Q: So that'll mean a total of four?
Q: Do you have your own lab at Netscape?
A: No, the machines we use sit in their server room. They have been
really good to us in terms of getting us what we need. They have a great
connection to the Net which we wouldn't have been able to afford on our own.
Q: Who do you interact with at Netscape?
A: The main person we've dealt with has been Bill Foss. He's been great
to work with.
Q: Do you guys have a remote access in -- telnet, etc.?
A: Yes, we do.
Q: Have you had any security problems?
A: Not that we know of. The only security problems are that Netscape has
their own firewalls and sometimes if they are tweaking their stuff, it
ends up interfering with our ability to get in.
Q: So the hardware you are running is 4 Indy's?
Q: And how much disk space do they have?
A: Not much...about a gig each.
Q: Does Yahoo only take about 30 meg?
A: Probably closer to 50 meg now because of the directory structure. The
database stuff is probably close to 30 meg. All combined, it's a total
of about 100 meg right now.
Q: When you are indexing, does Yahoo index in some way to optimize the
search? Is that all built in?
A: Yes, the indexing is optimized for our search. However, our search is
pretty primitive; it's essentially a grep.
Q: Have you looked at any of the third-party solutions out there?
A: Well, we certainly need to. What we have right now is not fast
enough. We are definitely looking at third-party search engines.
Q: Any in particular?
A: We are looking at Fulcrum, Verity, ....
Q: Illustra? We use that for our Sun Solutions Catalog. We've gotten
some pretty good results. You can do complex searches on complex
objects. They have a lot of Web stuff.
A: Basically, we want to find what works best for us.
Q: But you're only searching 30 megs, so it's not that big...
A: True. When you are doing a search of a relatively small text file,
it's not that big a deal.
Q: In our catalog, we have to accommodate 15,000 records that can be 10K
A: Right. It's definitely a different problem. They're indexing
gigabytes of data, which eventually we'll have to do also. For right now
though, we're searching on each entry's URL, title and comment, so
there's not that much data.
Q: How did you guys come up with the Yahoo categories? Did it just kind
of grow into what it is today, or is it based on something in particular?
A: Unfortunately it's not based on any existing categorization
schemes. When we started the number of categories was so small it wasn't
necessary to use a well structured scheme. If you go in science, there
are the obvious categories. We're not really specialists in any of those
areas, like biology. Basically what we have done is as categories fill
to a certain point, we break them into more specific subcategories. The
process has been fairly ad-hoc but is becoming more standardized.
Q: So have you been able to anticipate where the categories are going
A: We're getting better at that over time. For us ideally, we would like
to have maybe 30 entries at most in any one category, so that you don't
have to wade through a lot of stuff. So we always try to put a lot of
hierarchy in. You have to be real specific about what you're looking
for. We kind of look at it as complementary. You can browse it, and not
necessarily just browse it from the top level. But if you search for a
keyword, it will also tell you where in the hierarchy your request is
located. So, you can go into the hierarchy at that point and find
Q: There are other search web sites...have you guys worked with them at
A: Not really. But we know that their web crawlers go through our site
and index all of our pages.
Q: Is it true that Yahoo stays away from sex-based URLs?
A: Well, we don't have pornography. We did have requests from users
saying "please don't add this stuff" to the opposite response. But the
real reason that we actually took pornography out was because any time
we listed a site that had it, the next day - probably within a couple of
hours of listing it - it would be down.
Q: Too many people going there?
Q: Is that about the only thing that you don't put in Yahoo and do you
review links before you add them?
A: We pretty much add anything. The reason we stopped adding those
pornographic ones is because the sites would often be down and then our
links would always be dead. It would take awhile, and then users would
complain to us. We didn't really have time to sit there and look at the
links every day and see if they were still available. Plus just the fact
that those sites, in order for them to actually exist, can't really be
publicized in a public place like Yahoo.
A: So of course when Yahoo wasn't yet well known, this was fine. But
when we became a popular...
Q: Gotta be underground...
A: Yes, it's gotta be underground to even exist. With that exception i
we pretty much add anything that people request, provided that there is
some value to the content there. For example, there was one guy who
wanted us to place his hotel ad in several hundred categories. He's in a
town with many conventions, and for every convention that comes through
the town, he has a page for it that advertises his hotel. There's really
no information at all about the conferences themselves. So this is an
example of something that we declined to add.
Q: How many companies do you think are actually making money from being
listed on Yahoo? This seems like pretty good Web merchandising. Are
there companies that are making money off of Yahoo...by having a
listing, by being found?
A: We certainly get lots of letters from people saying that before they
were listed with us, they didn't get any requests, and after they got
listed with us, they started getting more requests. And so, it certainly
bumps up the visibility of certain companies. But it's hard to say how
much revenue is actually being generated.
Q: We heard that you guys were spending 15 to 20 hours a week updating
URLs while you were in school, and that now it's probably even more
time. Is this true, and how did you manage all the updating while you
were in school?
A: Well, we didn't have classes; we were working on our thesis. And our
advisor is out of the country on sabbatical.
Q: Big surprise when he gets back and sees what you guys have done...
A: He's been very supportive of what we've done. Now we're both on leave
and are devoting full time to the project.
Q: How long will you be on leave from Stanford?
A: Well, we just took a leave starting this last quarter in
April. Officially, it's two quarters that we'll be off. It kind of
depends on how things go.
Q: So your interest has shifted to the Web from EE?
A: Yes. This whole Internet thing is very intriguing. Plus, we're both
more practical than theoretical, so the Internet is a better fit in that
regard too. It's just a shame we couldn't combine the two so that we
could have finished school.
Q: Were you guys at Internet World?
Q: (Marc) At the show, I was approached by a guy who said "Where are the
Yahoo guys?" I didn't have the Yahoo information center shirt on or
anything. Your names have gotten incredible recognition. How are you
going to carry that?
A: One thing we have going for us right now is our name. We have gotten
a lot of good press. It's obviously a good thing and we need to take
advantage of that.
Q: You mentioned in one interview that you are from the old school, the
Internet. How do you view The Microsoft Network, how do you feel about it?
A: From our view it seems that they'll just be bringing a lot more users
to the Net. A larger audience for those providing services to the users.
Q: So it's a good thing?
A: We hope so. Especially for someone like us, that is, providing
directory services...their users will create a larger demand. Of course
Microsoft could compete with us if they decided to build Internet
Q: That's true. Is it just you and Jerry that do all the adding and
processing of everything?
A: Right now, it is. We're starting to look for other people to help out
Q: Do you guys have a nightly backup for your system?
Q: Do you guys have backup servers in case those ones go down?
A: No. It's not, right now, a very reliable system. In fact, they do
crash once in a while.
Q: So on the Internet, it works out well?
A: The Net as a whole is not that reliable, so our blips in service
don't cause major problems. However, we are certainly working on
improving our own reliability as well as those sites that we depend on.
Q: Do you guys have any mirror systems anywhere else in the U.S. or
A: Not yet, but we are looking into those issues.
Q: But not as a part of the business?
A: Well we've gotten a lot of requests in the past. People wanting to
put mirrors up. And one of the reasons we haven't done that to date is
because it can be a real headache, especially for material that changes
every day. It's not a trivial thing to do. And once it's out there,
there are a lot of headaches in order to change it later. So we want to
be careful on how we do that, and make sure that it's something that's
Q: After you find some engineering support in the various places?
Q: Especially with only the two of you right now?
A: We don't want to end up in a situation where there are different
Yahoo versions out there. Where some mirror machines don't have the same
data, that type of thing.
Q: There's a lot of synchronization to figure out?
A: Yes, some synchronization is certainly required. Right now we're
looking at mirroring, mainly for international sites because their links
are so slow. Some links are slow in the U.S., but I think within the
U.S, there's really not a need to mirror. Basically we're on the
Internet backbone, so it doesn't really matter if you're out here or
Q: When you do mirroring, are you going to franchise Yahoo, like a McDonald's?
A: We don't know. It depends a lot on how we decide to deal with
international sites. There are interesting issues that come up dealing
with localization. Right now, what we have is very U.S.-centric.
Q: Like Sun has Sun SITEs throughout the world.
Q: Maybe teaming up with something like that? I mean, they have
international sites, and something like that would probably benefit both
Q: Did you guys have a lot of requests asking for things to be
translated, or did English seem to be pretty much the standard?
A: English is pretty much the standard. The main request for mirror
sites is performance. People have talked about doing German and Japanese
translations, so there's interest in that. But the main concern
is...well, countries like Australia, who end up paying for their long
distance connection, which can get very expensive.
Q: It would be worth it to set up a mirror site in Australia, and they
could probably fund it easily with what they save by having it local.
A: Yes, one of the things we need to do is to find a long-term
partner. Our presence there needs to be a permanent one so we need to
find a company that is going to be around for a while.
Q: So you're saying it's the bigger guys, as big as they are with the
Internet right now?
YAHOO AS A BUSINESS TODAY
Q: Are you guys venture capital funded?
Q: Was it difficult to get venture capital or did they come to you?
A: Most of them came to us because we had mentioned to several people
that we were looking for money. It's quite interesting because many
venutre capital funded firms haven't yet supported Internet
companies. So we went through pretty much the whole process without
having to fill out a business plan. They realized, too, that things
change quickly on the Internet.
Q: Are you happy with the amount of money that you have?
A: For now, yes. But we quickly realized that you need a lot of
money. Especially when you see what T3 line costs each month.
Q: How much of your T3 line do you actually use?
A: Right now, without graphics, we're only about 3 to 4 T1's and
probably about 4 or 5 megabits per second out of a 45 megabit connection
Q: So you have a lot of room to grow?
A: Somewhat. But, bandwidth is still an issue and lines are not cheap.
Q: Have you thought of turning Yahoo into a hub?
A: That's an option. One option is to move and get our own T3. Another
option is to buy services from one of the new startups that have
multiple T3's and are in business to provide such services for a fee to
other companies. It doesn't make sense to have every company running
T3's. In one central place they have a lot of bandwidth coming in and
sorting the documents. There's no cost benefit in doing this, but the
problem is that a remote site is hard to maintain. In this regard,
Stanford had a really nice testing machine to test things out. It's
Q: So you are currently not paying for any of the access?
A: No. It's a mutually positive relationship with Netscape.
Q: And Netscape is promoting the community.
A: One of the main advantages is that Netscape views using directory
services as being very important. Without these services, the Internet
becomes unattractive to users. Users will then all go to AOL or
Compuserve. So they are very interested in seeing independent directory
services continue to flourish, and they are helping us to make sure that
this does happen. We both have a really good relationship with each other.
Q: So your overhead is pretty low right now.
A: Yes. But, we need to start generating some reasonable revenue. We
could have done that already, and, one of the reasons that we didn't do
that is because we wanted to do it right. It's important that we didn't
want to mess up on this. Netscape gave us some breathing room to get
going and to organize ourselves.
Q: Did Stanford work with you pretty well?
A: Yes, although it was never an official project. They let us do
whatever we wanted to do. We were just using our own workstations.
Q: They had to see an increase in net usage.
A: They saw it, but it wasn't really significant in the
beginning. Towards the end, Yahoo started picking up a fairly
significant portion of the Internet. But, no one ever complained that it
slowed things down. Stanford was very supportive, and now we are being
supported by Netscape.
GROWING YAHOO AS A BUSINESS
Q: Have you guys been in the industry before?
Q: Ever run a bulletin board?
Q: Do you have mentors for business advice?
A: Well, we're starting to bring in some advisors right now. Up to date,
we handled it all ourselves, and we pretty much split up the responsibilities.
Q: Do you both have to agree before you move ahead on projects?
A: Yes. We're now at the point of bringing in business people. Things
are moving so fast. We know that we're currently in a position where
there isn't really any competition, and we can't afford to make any
mistakes. We both want to get lined up correctly.
Q: Do you ever wonder if you can keep control as the company grows?
A: Any time you have investors, there's always the possibility of loss
of control. Right now, we're in a good enough position because we have
day-to-day, hands-on control. But it's not going to stay that way. So
it's up to us to bring in the right people to manage and grow the company.
Q: Are there any other investors besides the venture capitalist?
A: Hopefully we won't need money in the short term. If things go really
well, we might need more money to get some other things going. And
that's when we might think about becoming incorporated. We did have
interest from some corporations who wanted to invest in us. In fact, a
couple of months ago, we had a few options. One was to go off and fund
it ourselves. Another option was to sell to either a large company or an
existing service provider. We had offers to do that. We could have done
it, but we didn't want to work for a big company. We also thought that
our services would be better if Yahoo remained an independent entity and
was not tied to a particular company.
Q: Is Yahoo incorporated?
A: We're not a corporation; There are four positions on the board of
directors, and we fill two of them.
Q: Are you guys equal partners?
A: Yes. On our business cards, we are both Chief Yahoos.
Q: Will you be hiring a president?
A: We'll definitely bring in a president and other upper management.
Q: How many people are you looking to hire?
A: Between 4 and 5 people on the maintenance and technical side. We're
also looking for some people on the business side and in sales. First
thought is to hire library science people who have a background in
libraries, cataloging and stuff - the official library system requirements.
Q: How many people overall?
A: A year from now, we should have between 10-15 people.
Q: Will there be major changes in Yahoo now that it's a company rather
than just a couple of guys from Stanford?
A: Yes, hopefully for the better. The service will improve. - better
interface, better features. Over the year, it's been growing
constantly. I don't think we've taken anything off of Yahoo. We haven't
really had the resources to do that. So there are all of these features
that we've thought about wanting to add - things that the users
requested, and we haven't been able to do them. So now that we have the
resources, and we are starting to look at Yahoo differently. Hopefully,
the users will see that it will be alot easier to use - it will be more
Q: Did your accesses go up when you became www.yahoo.com?
A: We thought they might go down. We were just two guys who did it in
their spare time at Stanford. There's no way you couldn't like
Yahoo. Now, when we became a .com, it looks like we are doing it for
commercial reasons. So we thought maybe the daily accesses would go
down, but it just stayed about the same. Everybody knows when you
change. When we get to the point where we actually start putting the
advertising in, I think we will start turning people off.
Q: Where do you want Yahoo to go in the next year? In the next 5 years?
A: Five years is ridiculous to even think about. I mean, in a year who
knows what's going to happen on the Internet.
Q: Are you going to be a full company in six months?
A: We would like to remain small. We're not interested in becoming a big
company. We want to grow it well and, in order to get there, we have to
build the whole infrastructure. We basically spent the last year on it
with no pay. One of our main motivations today is to see if we can do
it. To be able to look back at the last year of our lives and like what
we have accomplished. That's some of the motivation.
Q: (Marc) Pays off when you see it grow. Do you get excited when you
think about the fact that at any given moment - even when you're in bed
at 3 o'clock in the morning - that there's a thousand people hitting
A: No. Actually the bad thing is that when a thousand people hit our
site, we think about whether it will shut off, and we are sleeping.
FUTURE YAHOO BUSINESS MODEL
Q: Are you basically going to try to stay away from charging the user?
A: We definitely don't want to charge the user. Generally, everyone is
in agreement that Yahoo needs to be free. For additional services, this
model may change. But most services we would like to keep free, unless
we get to a point where the resources really require support that is so
significant that there's no way you can provide them without charging
Q: What about charging on the back end?
A: We could talk about charging the lister - especially for commercial
listings. We would probably still want to give a basic listing for free
to everyone. Maybe charge for additional press listings - logos or
things like that.
Q: Can you elaborate on the business model?
A: We'd follow a yellow page model. Everyone gets one free listing, but
additional listings would be extra. And more listings are better than
only one listing. You pay for the increased visibility of a longer
description or a logo. We could also start doing more advertising ideas,
like a sports page could be sponsored by Nike.
Q: Like the Prodigy model, or the Netscape model which has ads?
A: Yes, but we'd do it a little differently. Say for instance that
you're interested in camping. So you go through the camping stuff and
maybe the page is sponsored by Northface or by REI. Users would be more
interested having ads specifically tied to the entries.
Q: Listers would also probably be more willing to pay.
A: The other model we're looking into is to charge the providers. We
want to get into that. Providers like AOL have some need for directory
services for users.
Q: Do you think service providers would be pretty interested in that?
A: Yes, we got some interest from them. Another idea would be to
specialize. For example, have a family-oriented version that has a more
selective, specialized focus, such as having a business directory.
Q: They could even do on-demand versions. For example, in any
organization with many users. you could focus on versions for specific
user groups. People would probably like these customized versions.
Q: Will you license Yahoo to other companies?
A: If our affiliation with a particular company precludes our talking
with other companies, then we need to revisit our current arrangement.
DAVID AND JERRY
Q: Are you guys best friends?
A: Yes, we're good friends.
Q: Same interests?
A: Yes. We're in the same research group. We share an office.
Q: So was it all worth it? Is it a learning experience?
A: Yes, it's fun. Hopefully the money that we have will sustain us until
we are self sufficient. We actually could have been profitable months ago.
Q: Is money ever an issue? Do you consider yourself to be a struggling
A: Yes. We got some money from Sequoia, but we're not getting rich.
Q: A year ago, you also had an option to get $40,000 plus in ads...
A: Well, up until recently we were living on RA [Stanford Resident
Assistant] shifts which is barely enough to get by.
Q: Any complaints?
A: Actually, we've been very lucky in that we've had a lot of support
from several sources. There has been no lack of information resources
because a lot of what we've done has been driven by the users. Their
suggestions are helping us, pushing us, voluntarily giving us good
information. For us, the only bad thing is that we don't have the time
and, sometimes the resources, to make Yahoo better. It's really hard to
complain about anything that wasn't really under our control. We should
have been getting more resources along the way, but we didn't anticipate
the explosive popularity of Yahoo. Six months ago, things changed. We
knew that things were going in this direction, and we're hoping that
Yahoo will succeed.
Q: You kind of want to pinch yourself to see if this is really happening.
A: Yes. Right now, the number of users in the lab is pretty small
compared to what we think we're going to need short-term. That's one of
the big frustrations. The other thing is we've just been kind of keeping
up. We do what we've always been doing. I'm doing the Xfile mode
catalog, which is essentially an unsolved problem. A lot of people look
at us and think we are crazy or stupid, asking us why we are spending so
much time cataloging when we should really be in a lab. We also get lots
of advice on whether we are doing things the right way, using the right
system, using a GUI, and things like that.
Q: You did make up your own classification scheme.
A: Right. Everything basically makes some sense. Looking at all the
tools that we use, we really don't have the time or manpower to step
back and think if this is the right way to do it. We're just trying to
keep up with the demand of the users. Hopefully, we're not going down
the wrong path.
Q: Do people think it's good?
A: There's a need for this type of service out there. But is it just a
short term thing until something better comes along? Who knows. For now,
without the day-to-day cataloging that we do, Yahoo wouldn't stay in
demand. So experimenting in a lab isn't the answer for us.
Q: Do you have any time to surf the Web?
A: Not for pleasure, But, we surf enough just by checking out the added
Q: You check them out as you add them?
A: Yes. We definitely have to verify the category. The best case is
where somebody does it for us - when they add it, they cross-link
directly. So we check them out and we get to see what's out there. But,
right now, we're sometimes unable to keep up with the demand of adding
sites. Between the email that we get and the requests to be added, we
definitely need more people on board.
Q: So has Yahoo taken away your social life?
A: Pretty much, yes.
Q: What do you guys do for fun?
A: Lately, we haven't done anything. Last year we had the best snow
since I've been out here, so we did some skiing.
Q: We work with computers all day and then we go home and we have
computers there and we usually work with them until the early morning
A: I never actually owned a computer. But the problem is I never go home.
Q: So you're always at the lab?
A: I'm either there or in meetings somewhere. About a month ago, we took
off to play some golf.
Q: Are you surprised by your new found fame - when you see your name in
A: Yes. But, it's not anything different than what I used to do. So, I
don't see it as fame.
Q: Do you get more emails?
A: A little bit. When a story appears, you get more phone calls.
Q: Are you recognized on the Stanford campus?
A: We don't have much public visibility on campus. The Stanford daily
paper hasn't done a story about us.
Q: What are your favorite Web sites? Like Mark and I like the Fed Ex
packing page where you can track your packages.
A: That's cool. Any time I Fed Ex anything, I use it. I also like some
of the world news and sports pages.
Q: What do you think about the newspapers that are charging access fees?
A: Yes. I get the San Jose Mercury News, but I never really read it, so
it seems reasonable to pay for the net version. It would be nice if it
was free. It's going to be interesting to see what happens to
information in general on the Internet. Now information is basically
free in most places. You just pay for the distribution of it. I think it
will be interesting when it gets to the point where they charge a penny
for each clip.
Q: There are some sites that are already charging.
A: Yes, you set up an account, type in your password, all kinds of
stuff. When you think about all the accounts that are out there - just
think about a thousand newspapers. You don't want to open a thousand
accounts. A long as the accounts are really cheap and they charge only
for actual usage...
OTHER INTERNET PLAYERS
Q: Who do you view as the big [corporate] players in today's Internet market?
A: Well, you need the hardware vendors to get the Internet going. So you
need the Sun's, the Cisco's, and the right infrastructure.
Q: Which companies do you think are proactive in making the Internet a
value-add or in keeping it free?
A: I don't think things need to be free. If you are talking about the
network, the only way that it's going to be long-lasting is to make it a
real commercial service. And, in order to do that, things are going to
have to become more commercialized and relatively free of government
funding or distortions by private companies, institutions,
whatever. Certainly the big companies, like Sun, are providing services
to the net that determine the business model, which is the needed
part. As far as growing the net, Sun is one of the main players, like
with the World Cup and that type of thing. I don't know if you guys made
money or did it as public service.
"There's a lot of interest in XML remote procedure calls such as SOAP," Fallside s ays. "The Web community needs a fairly simple way of enabling peer-to-peer Web applications to communicate. Right now, communication is primarily between Web browsers and Web servers." -- http://www2.itworld.com/cma/ett_article_frame/0,2848,1_2612,00.html
This archive was generated by hypermail 2b29 : Sun Nov 26 2000 - 00:03:10 PST