Jeff Clavier's Software Only

The Facebook unplugged at Stanford ETL

I have attended last night's session of the Entrepreneurial Thought Leaders (ETL) program featuring Mark Zuckerberg, the Founder and CEO of the Facebook, and Jim Breyer, Accel's Managing Partner who is sitting on the board of the company after his firm led a $13M round of financing (at a rumored $100M-ish post-money valuation).
The Terman auditorium was close to full to hear these two share their stories and experiences about the Facebook and its stellar growth. I have written about it in the past right after their financing, and actually this post is one of the most visited on my blog. Also check out their TechCrunch profile (and do read the comments from Facebook users).

Before I get into the actual session, here is an update on their metrics – which are (increasingly) stunning:

• 5M+ registered users
• coverage of 45% of US colleges (a total of 2,000 - representing 8M students)
• 80% penetration among students of colleges that are on the platform
• 10th most visited Internet site in the US
• 5.5B page/views a month (230M page/views a day)
• 8.5M unique visitors
• signing 20,000 new users a day
• repeat usage: daily 70%, weekly 85%, monthly 93% – can you think of another site that sees 93% of its registered users coming back every month ?

Mark Zuckerberg started the Facebook during his sophomore year at Harvard, in order to have a directory of students on campus in which they could list their profile, their activities and essentially what is going on with their lives. The application became popular beyond Harvard, and other colleges started requesting it.

Mark said that he has not conceived the Facebook as a social network  – which is a community application, it is a directory that he considers a utility that students use in order to find information which is socially relevant. Product development is about optimizing the application on scale and user experience, and adding new functionality: for example, they just added the ability to upload an unlimited number of photos. Expectation is that about 10 to 15% of the people on the network will upload photos.

One of the early design trade-off was to partition users by schools, and limit the visibility of profiles at school level. The decision was made in order to make the environment safer for their users, especially since 30% of them list their cell phones in their profiles. This contributed to an increased privacy, and a relative limitation of the number of  “lurkers:“ who are crawling other sites like MySpace.

On the reason why the first generation of social networks failed (as compared to MySpace and Facebook), he thinks that they have not focused on providing a set of utilities to their audience, they were merely about creating connections. They also had issues of scalability that plagued their growth.

Mark’s team is younger than most entrepreneur teams (many are under 21), and was originally made up of some of his college friends. The company has since quadrupled, and now has 40+ employees. They had started hiring engineers after Mark had built the first version of the site in a couple of weeks (note that he is a psychology major) and worried about scaling it. At some point they got an office, and had to face the issue of organizing the company and its product development, particularly in order to start releasing features in parallel as opposed to serially. Mark learned then the difference between managing dorm friends and employees who had to serve a community, and advertisers. Employees can’t be only engineers: finance, business development, sales became necessary functions to add.

The Facebook is recruiting heavily (Stanford) grads and experienced staff. Their primary criteria is raw intelligence, and alignment with the company’s market and culture. They actually don’t insist on raw development talent – more sheer interest and ability to get things done.

On international expansion, Marc explains that Facebook has to go  beyond internationalization and figure out the cultural implications of moving into markets like Europe, China, Japan,… The strategic choice so far has been to develop a high school product in the short term.

The session then moved into a Q&A with the audience that you can find in the extended post.

In a very timely way, Bambi Francisco has made a video interview of Mark Zuckerberg, which provides this data point about the original investor in the Facebook:

In the fall of 2004 former-PayPal-CEO-turned-hedge-fund-manager-and-angel-investor Peter Thiel -- whose fund is up 60%-plus this year -- gave Zuckerberg $500,000 to buy servers to support the growing audience base at Facebook. Thiel told me it's one of the best, if not the single best, venture investments he's made.

The video of the session is now available on the ETL website.

John Battelle interviews Omid Kordestani

I recommend reading John’s interview of Google’s Omid Kordestani, who also spoke at the Web 2.0 conference (see SearchViews’ account). Omid was the first sales/business person to join Google in 1998, and runs sales and global business development for the search&advertising giant.

There are some interesting nuggets of information in this piece:

• [Kordestani]  has to manage relationships with agencies that want more control over their clients' campaigns and with publishing partners who see Google as a prime source of online revenue -- and a long-term threat to their media businesses.
• Once [Google] proved that the text ads on Google could be successful and not interfere with the search experience, then it really turned into a science. We applied auction theory to maximize value -- it was the best way to reach the right pricing, both for advertisers and for Google. And then we innovated by introducing the rate at which users actually click on the ads as a factor in placement on the page, and that was very, very useful in relevance.
•  Right after the [dot-com] collapse, people were cynical about online advertising. But we monitored our traffic and the click-through rate of the ads. It quickly became apparent that AdWords was working. The model was perfect because people were willing to try something out for $1,000 and see if they could get enough leads to justify additional spending.
•  The measurability of online advertising will extend broadly to all areas of media. You have companies spending billions of dollars on television. As more and more consumers adopt technologies like TiVo, I think you'll be able to have much more useful forms of advertising -- more targeted, more measurable, and with new pricing models. Just imagine if we made it possible for our advertisers to quickly publish relevant ads that could range from the local plumber on one end to Super Bowl commercials on the other.
• As a matter of practice, Google will work with advertisers in any way they want to work with us, whether directly, through a traditional agency, or through a specialty search-engine marketing firm. Google sees agencies as crucial. Agencies create and execute marketing strategies for their customers, and Google helps them execute those objectives.

As an increasing number of consumer Internet companies rely on advertising as the primary, and sometimes unique, monetization mechanism, it is important to understand how they did it. Their key words: traffic, reach, breadth, relevance, measurability, ease of use, new pricing models.

Even if Google is delivering stellar results, there are still many areas of opportunities that will require new targeting and delivery mechanisms – for example to efficiently advertise in podcast, videoblogs, short form videos that will proliferate on Video iPods and cell phones, etc. On that front, it is just the beginning.

Data points regarding Google AdSense versus Yahoo! Publisher Network

Search Engine RoundTable points to a thread on WebMasterWorld related to the effectiveness, measured in revenues, of Google AdSense versus Yahoo Publisher Network. Message 9 provides a good summary:

I run Adsense and YPN on the homepage of my main site interchangeably to test which is better (equal pageviews at the end of the day, same weights). Both ad networks are both on target, although Adsense ads show more "less sophisticated" adverts (no brand name, mostly mom and pop operations); while YPN shows brand names. Google's Adsense gives me double digit CTR, while YPN only gives me 1/10 of Adsense's CTR. At the end of the day, even if YPN gives me higher earnings per click, Google gives me better revenues. It seems that G's targeting is designed to show what ads fit the page AND what ads are most likely to get click (probably due to some historical data of my target audience or whatnot). On the other hand, YPN only shows what ads can possibly be targeted to the content, but not really designed to generate high click thrus.

This is a key question for most Web 2.0 companies relying on advertising as one of their main source of revenue. At the end of the day, the depth of the advertising network and the more effective relevancy algorithms of Google seem to be the main differentiator leading to a substantially better yield.

Categories and Players in Contextual Advertising

As the growing number of Web 2.0 rely on advertising as their main source of revenues, I read with interest a great summary published by SearchViews, on the four types of contextual advertising.

• Search-based Contextual: serves ads matching the content of a site/page as precisely as possible.
  Google AdSense and Yahoo Publisher Network are by far the two largest players, with AskJeeves having recently launched their own network.
• Channel-Based Contextual: serves ads largely based on the targets/demographics of the channel (site or portion of a site).
  Value Click is the largest player in the space.
• Behaviorally-Based Contextual: tracks users across a network of affiliated sites in order to build a profile, that is then used to target the user with relevant ads, as well as generate off-line sales leads.
  Tacoda and Blue Lithium are mentioned, and I would add Revenue Science.
• In-Line Advertising: these are the big, annoying, roll-over ads that pop-up if you get too close to them on certain site, or even are displayed on top of the text.
  Vibrant Media is one of the providers of these ads.

Worth a read.