Google's History in Advertising

Sheryl started by providing a little history on Google's approach to search advertising. She said that "before Google advertising was focused on the advertiser," but when Google was developing their advertising model they started by "focusing on providing useful information to users." In other words, they "wanted the ads to be just as useful as the search results."

Regardless, Sheryl gave a good example of sponsored results being more relevant that organic results, when she talked about searching for "Sunday NY Times Subscription." This is a query that implies a very clear user intent. In her example query, both the first organic link and first commercial link pointed to a web-page allowing her to get a subscription to the Sunday Times. However, the ad included a special 20% promotion that the NY Times was running.

Extend to Content Network

After successfully delivering ads with their search results, Sheryl explained that Google extended its advertising model to a network of non-Google content. However, they continued to try and ensure the ads were valuable for users by contextually matching the ads to the content. For example, if you are visiting a car site, then the AdSense ads should be about cars. I've written before about how vulnerable a source of revenue I think this is for Google.

What Next in Online Advertising?

Sheryl then highlighted two big changes to the way that individuals are consuming content, neither of which will be surprising to regular readers of Read/WriteWeb:

• Social networks play a great role in user's context
• Individuals going beyond consuming content to also creating it

She shared three key points for the future of online advertising:

• Advertisers need to get better at creating a 1:1 experience for their users. As an example, she cited the work that Cadillac did with their MyCadillac campaign.
• Advertisements need to continue increasing personalization. This was surprising to me, given that I don't believe Google has publicly announced any plans to incorporate behavioral targeting into their ad delivery system.
• Users are demanding the delivery of information to be an experience and advertising must respond to it, just like content needs to.

Conclusion

While I generally agree with all three of Sheryl's points about where online advertising is going, I think the the use of experience is a very loaded concept. It is important to point out that one of the things that ended up being so key for Google, and other search engine advertising programs, was that the ads were text.This made it much more efficient for ad agencies to create ranges of ads, ensuring that relevant ads were consistently delivered. In fact, in some cases the internal marketing department was able to create the text ads. To the extent that improving the experience requires rich media ads, it may be a very difficult process to personalize those ads - due to a more limited inventory.

What do you think about where the online ad business is heading?

This afternoon at Supernova thirteen companies selected by TechCrunch and Supernova, from 130 applicants, presented at the Connected Innovators Session. The companies presented to a panel of experts including Mike Arrington from TechCrunch, Josh Kopelman from First Round Capital and Julie Hanna Farris - a serial entrepreneur. The panel held all of their feedback until the end of all the presentations.

What follows are brief summaries of each of the presentations:

adap.tv

As everyone is aware, the way individuals are engaging with video content is changing from broadcast to a more interactive model. Adap.tv is focused on enabling an advertising model to work with this shift. The presentation included a demo where multiple contextual ads were pulled from their inventory of "over 1 million ads." The video was about a newly released movie and the ads were for renting that movie on Netflicks and purchasing it on Amazon. After the presentation I talked to Amir (CEO and Founder) and I discovered that the ads are primarily served from ad networks and affiliate networks. While this makes the size of their network less impressive, what is great is that they have found a creative way to display text ads in videos. As Google has proven, text ads do make it easier to tap into the 'long tail' of advertising.

AdaptiveBlue

AdaptiveBlue is a product that takes a little bit of time to grok, but I think it is ultimately on the way to bringing the semantic web to reality. What is great about both of AdaptiveBlue's products (Blue Organizer and Smart Link) is they have really done a great job of delivering value to each individual user, while also harvesting meta-data that their system can leverage for their other products. This ultimately creates a great network effect. [disclosure: CEO Alex Iskold is a R/WW writer]

Aggregate Knowledge

Aggregate Knowledge was probably the most mature company out of the group of connected innovators. They are currently powering discovery for over 50 million users each month. Paul Martino, CEO, referred to this as the "world's largest implicit social network." They do this for both editorial content and product recommendations.

CastTV

CastTV is focused on improving the experience of searching for video content. They started by doing a search on Google Video for "CSI TV" and then repeating the experience on their beta site. While obviously this was a rehearsed demo, the results were actually quite impressive. They explained that these results were because of their focus on two aspects that are very unique about video search:

1. Crawling and Indexing
2. Relevancy and Ranking Algorithms

Critical Metrics

Critical Metrics is a music recommendation service built around aggregating reviews from music sites across the web and automating the process of purchasing that music. You can easily purchase or stream the music from services like iTunes, Rhapsody and Real Player. This is a very interesting example of music hyperaggregation, similar to my review of Internet Video Hyperaggregation a few months ago.

Jangl

Jangl has developed a secure way for users to communicate withe each other in context. The goal of the site is to ultimately build 'the world's largest people directory.' They started working with dating sites like Match.com, enabling secure communications between potential matches. They expanded this to work with a number of social networks.

Pando Networks

Pando Networks originally was a P2P client that was a simple option to share large files between people. They've had over 8 million people download the client to share files. Today, they announced they are releasing a publishing platform for sharing video content. They will create a white-label solution for publishers who are interested in delivering a client for streaming their content. They launched today with Blip.tv, Rever, and NextNewNetworks.

SodaHead

SodaHead was founded by Jason Feffer, formerly the VP of Operations at MySpace. According to Jason, one of the most popular features on MySpace was the ability to conduct polls within your network. Therefore, after leaving MySpace he has focused on trying to create a great polling infrastructure. They have a destination site as well as a set of widgets.

Spock

Spock is a search engine focused on the vertical of finding people. It seems like Spock has been in a private beta for months. The demo was almost identical to the one they did at the Web 2.0 Expo. In the demo they searched for Venture Capitalists and Bloggers and returned some of the panelists (Mike Arrington and Josh Kopelman). While it certainly looks impressive, the challenge will be when individuals are slightly less famous then those they returned in today's demo.

Wize

Wize is focused on improving the product recommendation process. The are trying to establish a new score (Wize Rank) that aggregates multiple reviews. Wize positions this as the 'consumer reports rank of the future.'

ZapMeals (Fake Company)

The company was billed as an 'ebay for meals.' After the presentation, the audience was informed that one of the thirteen companies was a fake. The audience was then asked to vote which startup pitch was actually fake. The majority of the audience voted correctly for ZapMeals. I'm not sure how well this will translate in this post, but it was actually quite fun.

ZenZui

ZenZui has developed a unique approach to syndicating content on mobile phones. In addition, they have built a navigation system built on 'tiles', that allows the user to navigate exclusively built on the numeric keyboard. In other words, some numeric keys are used for navigation while others are used to select content. See Read/WriteWeb's earlier profile of ZenZui.

Zing

Zing has partnered with device manufacturers to 'un-tether their mobile devices.' In other words, they have an application that device manufactures can OEM to allow users of those devices to stream and download content (music and images) to their mobile devices.

Conclusion

The panel had a number of observations. The two that I found particularly interesting were:

• The number of companies that are focused on improving the experience of using a mobile device
• The number of companies focused on vertical search problems

Both of these are markets that are definitely heating up.

There also was an interesting discussion about which companies would be most likely to reach a "major liquidity event." Mike predicted that Aggregate Knowledge and Cast.tv will. Josh is an investor in Aggregate Knowledge, but he also pointed to Pando and Adap.tv. He noted that Adap.tv plays in a large and interesting space (and disclosed that he is an investor in another company in the space.) Julie liked Spock and Wize, because they "solve a real problem I have." Personally, I was most impressed with Adaptive Blue and Adapt.tv. However, the fun thing about seeing companies this early is that we'll get to see how these entrepreneurs adapt to the marketplace and their customers.

What do you think? Which companies look most promising to you?

• Scale and diversity are almost beyond comprehension
• Expectations and needs will continue to grow
• 20 to 25% of the queries we see today, we have never seen before

I found the third point quite amazing. I would think with the number of queries that Google processes, they would have seen a much higher percentage of the queries before.

Next Udi explained that there are three levels involved in trying to deliver relevant information back to users:

• Users and Queries
• Models
• Languages

Users and Queries

Udi gave some examples of Google's ability to understand the different between two very similar queries. For example, Google understands that 'GM' stands for 'General Motors', while 'GM foods' is actually 'genetically modified.' If you search for 'B&B AB', Google knows that is 'bed and breakfast in Alberta', while 'Ramstein AB' is 'Ramstein Airbase'.

Google also will recommend queries that may deliver better results. For example, if you query 'Types of dogs' it will give results, but also suggests 'breeds of dogs' as a better search.

He then explained that they still can't find all the answers. As a fun example, he said the query "Why Search is Hard" is actually a very difficult query for Google to parse.

Models

Next Udi reviewed some new Google search functionality, which while not live yet - will be soon. Apparently, Google is going to start trying additional queries based on certain user queries. For example, the query "How much does it cost for an exhaust system" will pull up results from "cost of an exhaust system." Beyond just removing certain general words, they are also interpreting the question as part of the model; for example the following two queries:

• ‘overhead view of bellagio pool' to 'bellagio pool pictures'
• ‘fedora 5 losing network connections' to 'fedora 5 network configuration’

Different Queries for Different Locations

Finally, Udi talked about how results need to be different when the query is conducted in different locations. For example, the query 'government' needs to return results about your countries' government. I haven't tried this in other countries, but here in San Francisco the first result is for the US Government.

He also reviewed a tool at Google.com.eg, which actually takes a query in another language, translates it to english, runs the query, and then returns the results in that language. You can actually view the page in that language. There are a whole suite of language tools Google seems to be leveraging at: http://www.google.com/language_tools

I'm surprised there aren't more copyright issues here, but I'm not a lawyer. For an example, here is the Read/WriteWeb homepage translated into Arabic. Apparently, when Udi was demoing this for Larry Page, he asked why the images weren't translating. Obviously, there is still work to be done, but it is quite amazing.

Conclusion

After listening to Udi's talk, I must agree that while I thought search was complex, I probably underestimated some of the areas of real difficulty. It is amazing to step back and think about how conceptually complex this is. It sheds new light on many of our challenging experiences around the AltSearchEngine's Day without Google.

• Clay Shirky (NYU)
• Rueben Steiger (Millions of Us)
• Raph Koster (Arae)

The brief description in the program was:

"Do most people really want to immersed in 3D virtual worlds? And what are the real business benefits of these massively multiplayer environments? This session will examine which activities will migrate to virtual environments and when physical forms will continue to dominate."

The panel started with each panelist responding to the question, 'Virtual Life or Virtual Hype?'. Clay and Raph began by emphasizing that the virtual world market today is primarily composed of individuals who are playing games. He indicated that while there is a lot of hype surrounding 3D internet and metaverse things, they are really just a small contingent of the virtual world market.

Pic: psd

Rueben explained that he believes it is very early in the first innings of a game that will probably go extra innings. Therefore, he sees the question as a non-issue.

Second Life

The conversation then transitioned quickly to a discussion about Second Life and Linden Labs as a non-game based virtual world. Specifically, there was a long debate about whether Second Life is actually growing or not and how sticky the environment is. Rueben framed the issue well, explaining that for the first 3.5 hours of using Second Life, the attrition rate is horrible. However, after 3.5 hours the attrition rate drops quickly. The reason is that right around then people meet someone. Linden Labs is working on how they can engineer an environment that accelerates new members meeting other avatars.

Clay pointed out that while this is true (members staying after 3.5 hours are loyal) he had to point out that the overall attrition rate is still 90% and so he remains skeptical. He emphasized his point by stating that "I never in my life bet against the user." He said there were two key questions for him:

1) Do the users who like it, like it a lot? Yes

2) Do a lot of users like it? No

General Framework

At this point, Jeff Clavier asked from the audience if we could move beyond just talking about Second Life. Reuben verbally proposed a matrix that I have sketched below. Along one axis is whether the environment is 2d or 3d. Along the other axis is whether the focus is social or entertainment. Second Life is a great example of a three dimensional social application. World of Warcraft is a great example of a three dimensional entertainment application. Along the two dimensional axis, the video game consoles are excellent examples of entertainment focused experiences. Along the two dimensional axis, Gaia Online is a good example. Incidentally, it seems like Rueben's firm has decided to focus on the social applications, as they have brought 1/3 of the Fortune 50 into Second Life and they're now expanding to focus on Gaia.

Conclusion

To wrap up, the panel again revisit the opening question of what was real and what was hype. The panel rallied around two themes: the younger generation and geography. Raph pointed out that under the age of 20, "anyone who is not a gamer is an aboration." According to his experiences, they actually move as groups from game to game - often sampling 20 to 30 games a year. Although, they tend to relate most to the first experiences. The panel also pointed out that due to a confluence of issues, Korea has multiple virtual games that are 1/2 billion dollar (US) businesses. These influences include:

• Financial crisis giving people time
• Fiber directly to the apartments in major cities
• Respect for gamers cultural (all the way back to Go)

The panel all agreed that looking at Korea as an example, as the 'young generation' grows up and infrastructure is built out, there should be just as vibrant a market in the US. In other words, the panel all seemed to agree that it was not hype. I'm curious what the astute readers of R/WW would believe? Virtual Life or Virtual Hype? Please leave your opinions in the comments below.

Right at the beginning of applying for my research fellowship at Carnegie Mellon University, I had the opportunity to interact with the Nobel Prize winning economist Herb Simon. One of the things he said that had a huge impact on me was that 'winning a nobel prize is easy, you just need to focus on the right questions to answer.' As I've moved from academia into the business community, the lesson has still been great guidance -- I always try to make sure I'm focused on solving the right problems.

Reflecting on the state of social web applications, I think as an industry we're still figuring out what questions should be asked. My presentation at Supernova was an attempt to frame some of those questions by establishing a 'first principle' for social web applications.

First Principles

First, it's important to review exactly what I'm talking about. 'First principles' are the building blocks on top of which complex solutions are built. A few good examples of first principles being used in larger solutions are boolean algebra as a basis for the microprocessor and airfoil as the basis for jet flight. Both of these fundamental theories were later applied to invent literally world changing solutions.

I believe there is an equally fundamental first principle of social web applications, which is:

For each ‘complete’ user interaction ...

the user must perceive they have received more value than the energy (and attention) expended to complete the task.

On one level this principle is very obvious, because basic human nature seems to prove it. People won't continue to engage in acts they believe are more expensive then the value they are receiving in return. What I think is important is that users must perceive the ratio of energy and attention spent versus value received to be 'profitable.' The ratio is actually more important then the absolute amount of value delivered. This is probably best explained by an example.

Example Twitter

Value

This year at SXSW I went from skeptic to passionate Twitter user. Before experiencing SXSW, I didn't get why people would want to broadcast very short messages about themselves to a community of followers. It honestly felt like a valueless experience and therefore I hadn't created a Twitter account. However, I went to SXSW and wanted to find the best places to drink Shiner Bock (err ... best panels to attend). Twitter did a great job of connecting me with other attendees. After experiencing the value at SXSW, I ended up coming up with lots of other ways that I could use Twitter. In each case, the energy to setup and complete the interaction was less then the energy and attention I was spending to complete it.

I personally have found 3 common uses from Twitter each that deliver value to me:

• Short messages to your friends “what are you doing?”
• Networking at events
• Self-promotion

Energy Expended

While I'm now a loyal Twitter user who really sees a lot of value in the application, what I think is important is how easy it is to start experiencing this value. Therefore, the perceived value delivered (in this case finding free beer) didn't have to be that high. There are other services which require significantly more energy and attention to use, and therefore based on the first principle they have to deliver a much higher perceived value.

While there are certain things you can do to lower the energy required, such as:

• integrating with Open ID to make account creation easier
• simplifying your interactions and
• opening up / integrating with other services (such as Facebook)

However, ultimately you need to focus on clearly delivering enough value to your users in each of their interactions. While there certainly are specialist you can turn to for help (we an HCI experts at my company), this needs to be the mission of everyone involved in product design and development.

Design Implications

Thinking about the implications of my proposed first principle , I think there are three types of interactions that we should focus on when creating social web applications. In each case, the interaction needs to be delivering more value to a user then the energy they are exerting. This can be achieved across three basic types of interactions:

1. New interactions users find valuable
2. Returning significantly more value to the user with the same interaction
3. Returning the same or slightly more value significantly more efficiently

Network Effect

It seems like the most common violation of this first principle is in the area of applications that require a network effect before they are successful. While the value may evolve as they site grows, each interaction needs to deliver value back to the user participating at that point of scaling.

Example: Wikipedia

Interestingly, I used to look at wikipedia as an example of a site where they seemed to violate this principle for those who create entries. However, an experience I recently had actually made me realize they do a good job creating a positive value exchange even for the act of creating new entries.

Recently I searched for 'behavioral targeting' in wikipedia and didn't find any entries. While I'm sure many had left at this point, I instead realized this was a great opportunity to contribute a piece to wikipedia and continue to work to establish myself as someone knowledgeable in the field of delivering targeted ads on the web. So I created an account with them and created the entry. This was six months ago, reviewing the history I see that 36 people have edited that entry including most recently (today 6/20) the director of behavioral targeting for Yahoo. Through this experience, I've realized they do indeed satisfy the first principle across all types of interactions.

Conclusion

So I opened this post talking about the need to be focused on the right questions. The question that I proposed we focus on at Supernova today based on the first principle laid out in this post is:

Are your interactions delivering enough value to your users?

The presentation was well received at Supernova, what do you think ... is this the right question to focus on? What do you think of my first principle? I'll try to monitor the comment stream below.