Geek confessions -- what's driving Web 2.0

My observations on Web2.0 are the cover story of this month's MIT Technology Review Deutschland.
Since there's never enough space to go into the finer points, here are some of the detailed answers that Gaurav Oberoi (on the right) and Chuck Groom of the start-up BillMonk sent me about what motivated him and his former Amazon buddy Chuck Groom to leave their jobs and set up an online bill-sharing service that works through a browser and mobile interface.

Gaurav's and Chuck's key take-away: It's a thriving ecosystem with small and quick players.
It's become a slot machine rather than a fancy roulette table, so the geeks can play and don't need to hire MBAs any longer.
(Thanks for the picture of the BillMonk data center!)

* What is Web 2.0?

From a practical point of view, Web 2.0 identifies the second coming
of the web - a return to when investors and consumers alike see the
web as a medium with great potential. Once again, we see a surge in
investment, and activity in online businesses. In this usage, Web 2.0
assumes a very pragmatic shape; it helps differentiate new
vs. old. E.g. Flickr is Web 2.0, but Ofoto is old-school 1.0.

But what are the characteristics of this new phenomenon? What sets
apart a Web 2.0 app from it's 1.0 ancestor? What makes this
interesting enough for so many to invest in this space?

I think there are 3 important points:

1. Rich UI

Browser standards, and browsers themselves have evolved to a point
where they offer developers a great deal of power in defining the user
experience. In Web 1.0, apart from basic HTML, we only had crudely
supported Javascript. But today, we have CSS, DHTML, asynchronous
calls to the web server (AJAX), and great new HTML layout features
like floating divs - the majority of which are supported by all the
major web-browsers.

What this means is that it is now possible to build an application
with a rich interface without having to go to the desktop. This is so
huge that I must repeat it: if you want a rich app, you no longer have
to stick to the desktop.

Why is this significant? Supporting desktop applications is a
pain. It's expensive, pushing out upgrades is hard, and getting users
to install something on their machines is near impossible in the first
place (how many casual visitors would install Flickr or BillMonk,
vs. the number who will create an account in 10 seconds?).

Furthermore as any good developer/designer knows, you may have a great
solution to a very important problem, but if your software is hard to
use, then users will won't touch it with a ten foot pole. So if you
needed a rich UI, and you couldn't afford the costs of a desktop app,
you just abandoned your idea because the technology wasn't there.

But with the flexibility and power of the new browsers, this hurdle
was removed, developers everywhere rejoiced with the realization
(which came in the form of AJAX), and Web 2.0 was kicked off.

2. Low-cost

Realizing an idea on the web today costs a lot less than it did back
during the dot-com boom. Why?

a) Hardware is cheap.
Most ideas can be implemented on old Pentium III boxes with a tad bit
more ram, and maybe a couple of fast disks. There is no need for
expensive CPUs when all you're doing is reading and serving images, or
tracking someone's calendar.

b) Great software is totally free (as in free beer).
For building a production system: Linux or *BSD operating system;
Apache web-server; PostgreSQL or MySQL database; Ruby, Python, PHP for
scripting; and javascript libraries for AJAX, useful UI widgets, and
routines. Frameworks like Ruby on Rails offer a great deal of plumbing,
and help simplify common tasks - and therefore greatly help speed up
development.

For scaling a large system: MemCacheD for caching; pound for
loading-balancing.

Development tools: Subversion or CVS for a code repository; MediaWiki
(or any other) for documentation; Trac or Bugzilla for
feature/bug/task tracking; IM, GoogleTalk, AOL, Skype for
communication; and various open source tools like Gimp and Inkscape
for tasks like image creation and editing.

In the lists above I've covered the gamut of what is needed for
running a website, and a development house behind it. I can say this
with certainty because we compiled the list based on our own
experience in running BillMonk. All of this is high quality production
software, is used by places like Google and Amazon, and it is *free*.

For kicks, Chuck and I did a back of the envelope calculation on how
much it would have cost us to go with a Microsoft based
solution. Developer's licenses, and server operating systems alone
would have exceeded $10,000 - a phenomenal sum compared to $0!

c) Development time is low
Scripting languages have evolved to a point where they are fast enough,
and powerful enough to build relatively complex systems which perform
remarkably well. They tend to be far less cumbersome to write in than
something like C++ and Java, and have a much shorter learning curve.

This dramatically lowers the technical barrier to entry for someone
interested in executing an idea, and reduces the cost of hiring
developers in general.

3. Approach

Given that UI flexibility allows for more interesting ideas to be
realized on the web, and given the low cost of doing so, many
have decided that the risk vs. reward equation evaluates in their
favor and are giving it their all.

This mentality which spawns from the above two points is a key
aspect of Web 2.0. There have always been a lot of crazy high-risk
ideas out there; a handful of these could turn into something great,
but it wasn't worth the risk of trying out till now.

When you see a new internet startup today, you immediately think Web
2.0. Why? Does it use RSS? Is it based on social-networking or
community involvement? Does it use AJAX, or is tagging a big part of
it? Who cares? Does that really define a Web 2.0 app?

I don't think so. By those standards, BillMonk, which some online folk
claim to be a Web 2.0 app, just barely fits the criteria (not that we
really care).

I think Web 2.0 embodies the idea of the small, low-cost startup,
solving a very specific, and perhaps high-risk, problem, and typically
leveraging new UI technologies to make the experience compelling
enough for people to actually use it. Many people reduce it to the
technologies involved, but they are only enablers, not the phenomenon
itself.

* What functionalities and also business opportunities do these
  technologies add to the Net as we know it?

The technologies help reduce the barrier to entry into this market -
which creates a mentality where entrepreneurs can try crazy ideas at
low cost and therefore low risk. This economic incentive to innovate
and enter the market results in lots of players, and lots of action.

This is great for consumers who get lots of wonderful freebies:
calendaring systems, photo sharing, finance management, etc. But it is
also great for investors who realize that some great businesses will
come out of this world-wide R&D lab, and all they have to do is
carefully watch and pick the right ones early on. 

Big companies also love this, because instead of funding every little
idea in house which would be impractical, they can pick and choose
from the flurry of activity outside their walls and then incorporate
it for cheap - especially if it comes with a loyal fan base and lots
of positive press.

The great thing about Web2.0 is that it's a lot of small players
moving very quickly to tackle simple problems using novel
interfaces. Small and fast. That's the key. Fast development cycles,
rapid iteration, and an intense focus on customer needs are the petri
dish out of which the most astonishing innovations can arise, and only
small teams can escape the inevitable slowdown of management and
integration overhead.

* What makes Billmonk a Web2.0 company?

BillMonk is a web and phone tool that solves a well-defined problem of
tracking friendly debts between friends in a user-centric
fashion. That means that we try as hard as possible to make things
easy for our users, we listen to our user feedback, and we rapidly
iterate to make it the best possible experience.

We are buzzword compliant (the site uses Ajax, Ruby on Rails, Linux,
etc) but, as I argue above, that probably means very little.

I think what really makes us a Web 2.0 company is that we've bought
into the approach and philosophy of staring a company now: it's
low-cost, with a potentially high upside, so let's try our relatively
high-risk (in the traditional sense of the word) idea and see where it goes.

* Has the way startups can be launched changed in this second phase of
  the Web? If so, how (funding amounts necessary; manpower necessary;
  VC, angel and big player such as Google, Yahoo or Amazon
  involvement).

Yes. You're seeing a lot of startups based on very little seed capital
that are primary driven by geeks rather than MBAs recruiting geeks. As
mentioned previously, the low cost of hardware, and of developing
solutions means, the costs are very low - as low as in the few
thousands not tens or hundreds of thousands.

Furthermore, in the time that most large companies plan and get a
project off the ground, a small development team already has a product
in live beta with several hundreds or thousands of users. The time
scales are completely different, even while retaining high quality.

Interestingly, this low cost has allowed a lot of companies to be
managed by the geeks, as they haven't had to bring in biz-dev folks
to manage and fund these ventures.

But why didn't this happen before? Partly due to technology, though
its arguable that someone could have done this back in 2000 - but I
also think it has to do with the much larger audience on the web, and
the new channels for reaching lots of people in expensively.

In the dot-com boom, startups had swathes of cash with which to host
ads on TV (remember the 2000 super bowl?). Today, you can simply turn
to a couple of big bloggers, and hope that they find your story
interesting in order to reach thousands of people (e.g. Michael
Arrington of TechCrunch).  Alternatively, with a relatively small
advertising budget, you can spend some time tweaking Google adwords
and be paying less than a dollar per conversion. Not bad at all
compared to hiring an expensive sales guy or managing a print/radio/tv
campaign.

So again, low cost is defining how companies are started in this new era.

* Can you briefly describe what you did at Amazon, what prompted you
  to leave to start your own company and what you learned at Amazon
  that you could bring to bear on developing BillMonk?

There are two of us, Chuck and Gaurav.

Chuck started on the Buyer Protection team. He worked on the A-to-z
Guarantee program for 3rd party merchant sales, and worked on ways to
reduce Amazon's exposure to fraud. In search of hard computer science
problems, he moved to the newly formed Item Metadata Services team to
help design the next-generation systems that could automate knowledge
collection about the world of products, e.g. to know that bicycles
tend to have two weeks and frame-materials like steel, aluminum, or
carbon fiber.

I, Gaurav worked for the search and browse team, and focused on the
categorization of items on the Amazon site. This turns out to be a
really, really hard (and therefore interesting) problem when you take
into account the sheer volume of data and requirements, e.g. "show all
men's shirts of brand X which are not Y and which are scheduled to be
shown on the site on date Z".

I have always wanted to start my own company, ever since I was a
teenager.  While at Amazon, I watched as the web entered its second
phase, and I knew that the time had come. I have always been the one
to discuss opportunities and ideas with my friends, and so I was lucky
to find a friend and neighbor like Chuck who does the same. We are
drinking buddies and we had often tossed around the "what-if" of
starting a company. After a lot of thought, we both decided to leave Amazon.
Chuck left first and I followed soon thereafter.

The main lessons of Amazon are:
1) Listen to the customer. Obsess over the customer. A happy customer
   tells friends to use you. A happy customer suggests things that will
   make other customers happy. When you make a mistake, apologize; that's
   usually what users want most to hear when unhappy.
2) Scale. Plan for growth, identify bottlenecks before they manifest.
3) Release early, release often. Break big changes into lots of stages.
4) Process. Track your bugs and features. Create simple timelines.
   Write a one or two pager design doc and discuss the idea before spending
   several days going in the wrong direction. Funnily enough, this was
   something I learned from my team - Amazon in general is not good at this
   (though you might not want to print that).

* What did it take to get your company off the ground in terms of
  staff, programming time and financing?

It was surprisingly easy to get started. The two of us did everything.
We started in August and released in January. In that timespan, we
incorporated, filed a provisional patent, set up our production
machines and network, iterated on UI mockups, then wrote the BillMonk
application and support tools.

The company is self-funded, with even ownership. While we both have
savings, we're not awash in gobs of cash. We've had to be extremely
frugal about how we've gone about this idea. Being frugal is a great
exercise for any company, I think. Since we're a tech company, all
reduced all our costs to just three things: filing company paperwork,
computers, and internet bandwidth.

The server room is Chuck's closet.

* Do you see a new bubble building around Web2.0?

I certainly see a hype bubble, and I think that some of the Web2.0
companies that were bought out by Yahoo! or Google are probably
overvalued.  However, what we don't see is a ton of venture capital
flowing into Web2.0 companies. A lot of them receive modest venture
capital, but not enough to, say, rent entire floors of high-rises in
San Francisco and fill them with Aeron chairs. So, if there is a
bubble, it's largely restricted to the bidding wars between tech
giants for aquisitions of already popular sites and not in speculation
on budding ventures.

* Are there too many mee-too companies in spaces such as social
  networking, bookmarking, tagging etc. or is it just a healthy,
  thriving ecosystem we're looking at?

I think that a lot of what we're seeing is half-business, half geeks
playing around. It's always good for geeks to play, even if there's
often not much of a business plan. However, I think that we have to
stay grounded in the reality of actual usage. These companies only
matter if they're used by people outside of the blogosphere -- that is
to say, if they come to be identified for the problem they solve
rather than the buzzword bingo of approach. I do worry that the
blogosphere is sometimes a bit too navel-gazing and forgets that its
bubble is not the entirety of the Internet's ecosystem.

To put it simply, does the average college student even use RSS?  Does
he care if there is yet another site which aggregates news from his
favorite blogs? Or wait, does he even read blogs?

But that is being too harsh. I like your term of thriving ecosystem,
but just like natural selection kills of weak species, I think a lot
of these ideas will die off because the audience for them is too
specialized - and perhaps they are ahead of their time.

* Has the rise of all these Web2.0 companies changed the importance of
  Silicon Valley as the hub of Net innovation -- i.e. less need to be
  here. You are in Seattle after all).

Actually, no; I think the Valley is still the center of it all. We
were reminded of this ... when we flew down to Stanford for
Entrepenuer 27. We were surrounded by like-minded geeks, and the buzz
of ideas was unlike anything we had found in Seattle. While we don't
need to be in Silicon Valley, it would make basic networking a lot
easier.