Barney Pell's Weblog

o Andreas Stavropulos, DFJ
4Info
Evdb
Technorati
Akimbo
H5

I can't think of any new economy consumer business that became successful that did
not center on search, period. The internet unleadshed all this info,
transactions, and floodgates opened. The immediate problem is what about
transaction, information, coordation costs.

Niche, vertical, horizontal. Look at the biz model.
1. volume: how many consumers can you get through your engine
2. value: how much can you claim for each action that comes through
You don't need both.
H5 going to business litigation, risk management, compliance type, more
enterprise than consumer. From an eyeballs perspective it's not huge.
But the value per rseult that comes back, value per page (.20-.50) is
huge. So they went from 0 to 18M in bookings in 6 months!

Seasrch opportunities on new devices. Eg.g 4Info locally relevant search on
mobile devices. Become the new "eyeball bandit" (to quote SJ).
Web2.0: on board ot technorati, invested in EVDB. How to manage distributed
real-time info if you can't assume that static html pages are where the
content is in the future. Need to think about rss, pings, podcasting,
blogging, things very real-time relevant that lose value quickly over time.

o Theresla Ranzetta, Accel Partners
social-networking / search
We think about focused media property. There is some search element, start
with usually directory, maybe tech as well. But add human eleemnt for the
filtering capability, vs relying purely on tech and algs. Better is when
human element can come not just from editors on staf, but from end users
itself.

examples:
Thefacebook: Originally a directory for college students. But the real power
and value, why they have 65% of users log on once per day, hours total are
unbelievable. Because it has a real community that mirrors a physical
community, and users contribute not just content and profiles but do their
own organization, introduce their own groups, etc. WIth a very clear and
interesting demographic (college age eyeballs). The early days of search was
a directory or ontology with humans who help add value. As the web got huge,
you can't do that with humans. The beauty of a focused property is you can
actually scale that. Now with web2.0 you can use your tools to pull in
organiation.

Was at a conference recently. PT had very tiny reach, vs cbs news. But
advertising revenue for PT was 5x the revenue. If it's just a niche
technology, not big deal. But also a very rabit loyal user base that comes back time and
again. The thing you can't replicate is the loyal audience that comes back to your
site again and again.

Brightcove
Slammedia

o Jeff: What kind of matrix / threshold do I need to reach to get you
interested in my vertical play? How much do I need to build first, when you
you take the risk?

Mark: We like to see early traction (Yahoo had thousands, Google had tens of
thousands), because you never know what's going to take off. We've seen a
lot of companies come in that show some traction. When we made the first
investment in LinkedIn, they had 30,000 members (now 3 million). When we
invested in Paypal there was 1 guy, a year away from product, so there are
outliers. But by and large we like to see them started at least so we can
see what the metrics are. Call it 10,000 users, enough scale so it's not 5
guys and a dog.

I've seen many vertical search companies that shouldn't be VC financed. I
can build a nice $20-30M company. Not a VC company as we need $100M
threshold. But for example hotornot, they did very well without VC funding.
We had one VS come in, 100% of their target market had to go to the site
5x/week in order for it to get close to a 100M company. That's not going to
happen.

Andreas: For consumer facing things, that makes sense. But actually not that
hard to make some assumptions on a piece of paper, what do you really
believe, and do breakeven analysis for typical behavior.
e.g. can I monetize the typical user to the tune of 4 cents per week? That's
a more manageable number.
I'm aamazed how people get enamored with new algorithms, AI, neural nets,
but don't step back and do simple math (multiply and divide) to see what do
I have to believe about usage for this to be a billion dollar market?
Those tend to be pretty easy discussions to figure out.

o Jeff: In new environment, with cheap hw and bandwidth, what does it take
to build a search engine that will reach scale? Some people are
bootstrapping their companies with $50K. How much do you plan to put in some
of those companies?

Theresia: When you're raising funds, it should be very capital
efficient. Mark Andreesen spoke about $100K or less.
After that the amount of capital depends on:
overall size of market opportunity in the near term (12-36 mos post
financing).
You can't acquire customers baseed on money, but if you've proved customers,
you can accelerate growth with capital.

Mark: Our investment in Yahoo was $750K. The face book got very big before
our round, just 4 guys programming and drinking a lot of red bull. You can
get far on 10s of thousands.

Chris: If you get 4-5 very smart engineers who can go without salary for 6
months, VCs may be irrelevant. Entrepreneurs can make a good living selling
search companies to the big players for $20M, who are in a feature war competing with
each other. With adsense you can have a 100,000 person sales force working
for you on day one. SO if you can afford to make that 100-200K investment,
there is a lot you can do.
Take the money when you really want to scale either the engineering team, or
the sheer marketing scale that blows the thing out on the traffic site.
I would be selfish if I were you, and think intelligently about when and if
to come to VCs. There is a lot of moeny to be made from sole proprietors.

Jeff: The besrt time to raise money is when you don't need it.

q: How important is growth rate, from investor perspective?

Mark: A friend of mine has 3 guys doing 1.5M per year. But not a VC growth
rate. We need to see a company get get to $100M in 6-8 years, taking not
more than 10-15% of the market (as a rule of thumb).

q: On pre-VC stage, how do VCs relate to angels?

Mark: I made a $100K investment last week.

Andreas: All our funds are getting smaller. Only $400M. For technorati we
didn't even do a round, just participated in the bridge. We do sub-1M
investments all the time to build a relationship for the professional
round.

Mark: It's a sellers market. Sometimes a firm will work with VC, sometimes
with Angels.

Chris: There's a lot of $15-30M acquisitions going on out there. We
incubated Oodle, Fatlens with a few hundred thousand $ as we thought they
could be enormous. There is a handfull of small and focused funds that will
make good returns as they can make money on the $20M acquisitions. We can't.

Jeff: With Oodle and Fatlens, were those already known entrepreneurs?

Chris: Yes, when you know the folks, a powerpoint and belief you can deliver
order of magnitude better experience than alternatives out there is enough.
Others like Myspace, you could have never predicted that was the mix that
would sign up 75K users per day, so we invested late in folks that are not
our typical entrepreneuers.

q: Craig Kaplan. Other monetization models than ads, eg. paid search like google
answers?

Theresla: Ads, transactions, and (much lesser degree)
subscriptions. Vertical search sites with largest exits had a combo of 2 of
those. For shopping.com, shopzilla, 50% paid search, 50% transactions.
We want to see an interesting user experience people are using. We don't
expect you've figured out every last way you're going to make money from
your concentrated user based. The best businesses leave themsevles open to
multiple ways to monetize their user base.

Andreas: Once a result is delivered, people are doing it for a reason. To
the extent these actions can be valued in a market or auction like
environment, where you can expose the info and let people bid them up to
their true market value, that's very powerful.
H5 is just processing pages for a particular vertical. The real question is
what is the value that ensues after results come back.
For H5 it could be turning a law firm on its head, that's a very high value.

As for expert advice, we were involved with that years ago, the big issue is
it's hard to prove the value of the result I'm going to be getting before I
get it. So how to avoid taking a huge discount on the market based on the
uncertainty of the result itself. Our guys switched the model to find a way
around that problem.

q: Eric Burns, StarsGuide, another couple of stanford guys starting a search
company. Are you really looking for small early stage investments? Or are
other angel funds savvy to the movement of vertical search?

Mark: All of us would be interested, as well as angels. We all do small
investments in this market which we believe in strongly.

q: Jeff: What does it take to avoid flipping.

Andreas: signals are clear: VCs only do the small investments if we think it
will lead to something huge. The limited resource is time and attention. We
can only do so many deals in a fund. If you do a $100,000, you only think
it's giong to justify the time and attention.

Chris: If it can be a $1B company but only takes $100K investment, we'll do
those all day long.

q: Jeff: When you do a 100K investment, how much time does it take?

Theresla: We don't think about $$ in, we think about investment in time.

Mark: We never use the phrase "put money to work". Not ROI, but ROT: return
on Time. That's the resource we have.
A VC can only do 1, maybe 2 early early stage companies at a time.
The average VC firm does 15-18 investments to a year. Each partner maybe
does 1-2. So if I invest in you, you're my investment for the year, so I
have to believe that thingi s a big opportunity.

q: James Healy, started Worldblog that will search blogs, but also track all
user-geenrated content. We're ex monster.com and collabnet people.
Mark mentioned 30,000 users as a point to speak to VCs. Are those
registered users?

Theresia: Not minimums, just prove to us it's more than just your friends
and family. Registered users can be more than regular visitors. But if
you know the same people are coming back again and again, that's what I'm
looking for. I certainly don't represent the vast majority of internet
users, so I can't use my intuition about how compelling this will be.

Mark: It's not just users, but also usage. How often are the users come
back.

q: Evan. How to judge a comptitive aspect. How to gauge barriers to entry?

Chris: Distribution and traffic, and a strategy for how to crack that nut is
not trivial. The big guys aren't sleeping anymore, they're paying attention.
With order of magnitude better user experience, people will find you. There
are many ways to see the market, just do some search spam, people will find
you and you'll be on your way. Really creative thinking on how you get
distribution. When Myspace goes to 18M users with no marketing cost, it
shows it can happen. Those are competencies you have to have on the team.

Andreas: The big guys aren't winning because of their better search
results. Google's going after MSFT. It's hard to fight on multiple
fronts. Large companies are much harder and slower about reacting to new
threats. By the time they figure it out, it's too late, you've created
value. So there's no reason why these guys won't suffer the same thing that
they took advantage of as small companies.

q: As vertical search companies proliferate, there is bound to be
duplication of efforts. Do you see strategic technologies?

Jeff: That will be addressed by Barney's panel this afternoon.

Mark: Yes. Advertising services, we're investors in Adbright. A couple others
we're looking at.

Chris: Talk to efficient frontier on how to use paid search to get audience
to your sites.