Tuesday, April 26, 2011

An example of prediction markets

For the British royal wedding of Prince William and Kate Middleton, here's one of the most interesting odds listed at Bodog.com
Royal Wedding 2011 - Which of these guests will be seen crying during the ceremony? Tears must be visible for winner
Carole Middleton (Mother of Kate)3/2
Philippa Middleton (Sister of Kate)5/2
Michael Middleton (Father of Kate)10/1
The Queen12/1
Prince Charles25/1
Prince Harry25/1
Prince Philip25/1

Prediction markets

Prediction markets are only effective if a lot of people participate.  What are the best ways to encourage more traders and trading within internal company prediction markets?

As Cowgill suggested, we should combine financial rewards (cash, gift certificates etc.) with other incentives: social rewards, processes, infrastructure, and even perception.

Social rewards play a key role here since usually play money is used instead of real money and financial incentives are very decent. For these reasons, we should not make it anonymous. Everyone should know other players so that they have incentives to play better and more seriously. People are as competitive and caring their "faces" in games as in work and life situations. 

For processes, a market and its outcomes should be selected so that most, if not all, people are interested in the market and have some uncertainty about the outcomes. This would attract more people like Dolores Haze, who stated "It didn't tell me anything I didn't know; it just confirmed my intuition."

Infrastructure can be very broad but some specific examples are: 1) make the user interface very simple and at the same time attractive and 2) provide various tools and data for each market and its outcomes to help users get started and do further research.

Perception of prediction market participation is also important. Some people may be concerned that if they participate very actively, their co-workers may think they are not doing their "real" jobs. Therefore, specify the maximum time they can spend at work for this activity and then use their non-work time. This can discourage some people from participating but then help many more people participate actively and confidently.

Tuesday, April 19, 2011


What are the similarities and differences between a community-driven product development process and a traditional product development process within a firm?

- Both go through basic development phases such as design, production, marketing, distribution, and customer services (though the community-driven one engages community tremendously in several phases, especially design and marketing).
- Both try to satisfy customer needs, either via formal market research as in traditional process or through community voting as in community-driven product.
- Both need to find more ways to bring in new customers such as rewarding referrals (though the community-driven one has users refer to others through links on their own blogs or via social networks instead of referring to others traditionally in person or via phone).
- Community-driven product development process engages a great variety of activities performed by the community as in the Threadless case: submitting designs, critiquing submitted designs, blogging about their daily lives, posting songs and videos inspired by the designs and most importantly purchasing t-shirts. In contrast, a traditional one is typically constrained by activities performed by the firm itself and involves no or limited community activities. For example, the latter has a team of designers and after some market research, comes up with new designs, and picks a few that it considers the best itself. The only community activity for the latter is to purchase whatever company designers (not "community" designers) make.
- A website for community-driven products is used not only for selling products but also for varying other services to meet different needs of the community. As in Threadless, people go to the website to buy t-shirts, to use the site to practice design skills and get feedback from fellow artists, to try to win design prizes, to interact with other consumers and designers, and to participate in blogs. A website for traditionally developed products is mainly for selling and might go as far as allowing buyers to comment and vote on products.
- For marketing, a traditional process pays for ads on public mass media such as TV, magazines, or radio. Recently, these companies also pay for ads on the Internet. For community-driven ones such as Threadless, they exploit less expensive and arguably more effective methods such as word-of-mouth marketing and press coverage as "the two major sources of Threadless community expansion."

Tuesday, April 12, 2011


LinkedIn recently hit 100M users. If you were in charge at LinkedIn, what would be your strategic goals for the next several years?  How would you achieve them?

LinkedIn has been doing great and in order to keep the momentum, it needs to stay ahead with some strategic goals for the next several years.

First, it should attract more professionals by focusing on what it has done well, PNS and by providing more relevant services such as for a certain skill or career, what degrees, certifications, or tests you need to consider. It can try to add some SNS features but there are 3 cautions. First, it can start just some basic features to see how users react. If it receives encouraging feedback, it can expand more. Otherwise, if it experiences an overall negative feeling, then stop. Second, when introducing SNS, it should be separate from PNS. It should provide a separate tab that directs users to a different page where users in the PNS services have to obtain mutual acceptance before becoming connected to others.

Second, it should open up its platform to third parties to bring in more services, fresher ideas, and ultimately, more users. It is just a matter of time since this is a trend and it proves beneficial to companies. Obviously, it needs to take cautious steps but should not avoid that. It needs to control the quality and the relevance of services. Also, privacy is crucial to build and maintain trust so a well-balanced strategy of openness and security is required.

Finally, it should partner with other companies that provide valuable services. It has already worked with several ones but it needs to explore more. For example, resume critiques, image building, and rental services. Based on the popularity (how many users use a service and how frequently they use etc.), it can list services in order of popularity, innovation, and relevance.

Wednesday, April 6, 2011


How do Wikipedia’s processes for creating and modifying articles ever lead to high-quality results? In other words, since anyone can easily edit Wikipedia, how is it that good (and usually accurate) content emerges?

As the top-ranked science journal Nature compared the accuracy of science entries in Wikipedia and online version of Encyclopedia Britannica on the same set of 42 science articles in late 2005, the former had 162 errors while the latter had 123. This is quite interesting given the fact that Wikipedia content is generated by the public and free while the other is considered the hallmark for factual authority and very expensive.

There are several reasons behind the relatively accurate content in Wikipedia. First, most people who dedicate their time and effort to creating and/or editing have some knowledge in the subject they write about and more importantly, are interested in sharing knowledge with others. Second, people, by nature, feel good when they voluntarily share something with others. Third, each person's knowledge is limited but if people contribute, the content will get richer, more comprehensive, and more accurate. Finally, if people make mistakes or intentionally put incorrect or false information, all it takes to delete or revert to an older version with better content is less than a few seconds, just hitting a button. This concept has been leveraged in several commercial tools such as Confluence that helps thousands of organization work collaboratively and more productively.

Another person got fired because of his Twitter comments

From CNN:
"The six-figure voice-over job with benefits opened up after Aflac fired comedian Gilbert Gottfried for his fowl, er foul, Twitter comments regarding the Japan earthquake. The comedian later apologized."

Tuesday, March 29, 2011

Social media

In general, how should organizations deal with issues posed by user-generated content and other information spread over social media?

Today, social media has significant impact on organizations. One individual transaction with one customer can be known to millions of people around the world and change organizations in a positive or negative way in several aspects: reputation, finance, and even their own organization cultures.

Organizations should deal with these issues in various forefronts:
- Be aware of the power of user-generated content and information spread over social media. It's hard to accept the fact that organizations do not completely control messages that are delivered to the public anymore.
- Overhaul organization cultures for internal deficiencies or weaknesses and improve or transform organization cultures. Sooner or later, they will need to provide a place to have direct 2-way contact with customers and when time comes, they can mitigate risks of getting too many negative feedback or backfire.
- Provide a means via social application technologies for customers and employees to get involved in organization activities: from ideas to implementation to feedback.
- Have resources dedicated to the social application and be proactive in delivering organization messages via announcements, blogs, and programs. When issues emerge, resolve as quickly and appropriately as possible.
- Be creative in getting customers involved and tolerant with necessary changes both external and internal.

Tuesday, March 22, 2011


How do network effects (cross-side and same-side) impact Hulu’s business model?

Hulu has three major groups of customers: content owners, users, and advertisers, who have strong network effects towards one another.

Content owners:
There's a Penguin Problem here since this is a relatively new market and therefore, the uncertainty is high and each individual content owner has their own long-term expectations. 
No-one moves unless everyone moves. 
Same-side impact
- The more content owners, especially those of popular content, join Hulu, the more likely others will join. The more owners join, the better for Hulu brand. That's why Hulu invested in a "long courting process" for Disney to join and offered Disney a stake of about 30%, similar to that of the founding partners because Disney was a reluctant but very influential player that can bring significant benefits to Hulu network.
- The more content owners join Hulu, the less power they have on Hulu since Hulu will not depend on just a few owners.
Cross-side impact: 
- If owners of popular content do not join the network, fewer users will visit Hulu and the network growth will stall and may drive Hulu to failures. On the contrary, with many popular content providers, Hulu can attract a large use base.
- With many popular content providers and bigger user bases, advertisers are more willing to pay for Hulu.
- If more content owners give exclusive licenses to Hulu, they will create more values for users and advertisers, which helps Hulu attract and retain more users and gain more advertising revenues.

Users play a crucial role in Hulu's successes.
Same-side impact
- More users can create more values to other users by helping Hulu learn more about users and give more relevant content and services, which helps Hulu.
- More users help one another through discussion boards, ratings and reviews, which also helps Hulu.
- More users can make network slower, which has negative impact on other users and Hulu.
Cross-side impact
- More users create more values for content owners and advertisers.

Similar to content owners, there's a challenge of uncertainty felt by advertisers.
Same-side impact
- More advertisers will bring in more money for Hulu. Furthermore, more advertisers will increase competition among one another and Hulu can raise the rate and earn more advertising revenues.
Cross-side impact: 
- More advertisers will attract more content owners to join, which helps Hulu succeed.
- More advertisers may annoy users and drive them away, which hurts Hulu.

Wednesday, March 16, 2011

Studios' and Netflix's Newest Threat: $1 Online Rentals From Zediva

The company thinks about its online service as a DVD rental and remote DVD playing and viewing service rather than a streaming service or anything else. An interesting combination of "low price point a la Redbox with place-shifting technology a la Slingbox".

Sunday, March 6, 2011


What were the key factors behind Google’s early success?
- If you have only one word to describe Google, what's that word? Google it!
At the beginning, it specialized in just one thing and did it very well: algorithmic search. Google.com offered more relevant and reliable search results with its index of one billion web pages, was revolutionary in its extremely simple interface, did not lure users to stay as long as possible on its website, and initially carried no ads. As it succeeded in this search service and built the customer trust and brand name, it began expanding other services: emails, maps, calendar and so on.
- Another day? Then another Google
Google constantly changes: It has learned from others, continuously improved, and frequently delivered innovative services and products. This has been a key factor not only for its early success but also for its sustainable growth. Examples are abundant: it learned from Overture but used CTR instead of CPC; it learned from DoubleClick and expanded AdSense; it learned from several email providers (e.g., Microsoft, Yahoo!, and AOL) and came up with Gmail. It can achieve the constant changes thanks to its obsession with innovation through adopting unconventional approaches for managing innovation.
- IPO? Yes and No
Yes, it did go ahead with IPO, but No, it was not the conventional IPO prospectus. The dual-class equity allows Google's management trio to have more power to be innovative and take risks without external pressure about replacement by investors. Ultimately, this dual-class equity protects the corporate values: 1. don't be evil; 2. technology matters; and 3. we make our own rules.

Thursday, February 24, 2011

Challenges facing Internet startups

Based on the Taxi Magic presentation and the MusicJuice.net case for next week, what do you think are the key challenges facing Internet startups as compared to brick-and-mortar companies?

Obviously, brick-and-mortar companies encounter several obstacles that Internet startups do not, but here, we discuss the other way around: the key challenges that Internet startups face compared to brick-and-mortar companies:
1. Harder to build trust
When people open a brick-and-mortar company, customers come and feel more secure doing business with them because their presence is known to several parties: property manager/landlord, county/city tax department, distributors, and so on. From customers' perceptions, brick-and-mortar companies also tend to stay long-term since it takes time to move away with  furniture and goods/products. On the contrary, customers tend to be more suspicious when browsing a website because the website can disappear in an hour without leaving any traces and customers have no way to track down if they lose money, need to return, or simply contact the company. As an example, Tim Csontos said they initially asked customers to give credit card info and no single person was willing to do that. If it had happened at a brick-and-mortar store, the chances that customers accepted that request would have been higher. Similarly, among the feedback from MusicJuice was "I do not trust the website with my money." This trust issue can have domino effect: I don't trust it, then my families and friends don't trust it, either.
2. More competitors
Brick-and-mortar companies do have competitors but usually within a specific region, there are no or just a few competitors. In contrast, Internet startups usually encounter more competitors due to:
- Internet is borderless, so competitors can be all over the world. For MusicJuice based in Toronto: competitors Sellaband from Amsterdam and Slicethepie from Berkshire.
- Internet businesses can be easier to be copied. Brick-and-mortar companies usually involve customer services and harder to imitate. For internet startups, if their business models are good enough, giant companies can copy and with their brand names and huge resources, they can beat those startups.
3. Unable to reach customers
Brick-and-mortar companies usually target on local people, So, they usually pick a great location with high traffic or in a shopping mall, so they get more exposure to local people. They can drop leaflets at residential homes to invite them to visit. For Internet startups, they need to deploy different marketing strategies. Taxi Magic resorted to well-known "tech people" such as the co-founder of Twitter and a person in San Francisco who has more than 1 million followers on her blog. It also managed to get voted among the top choices on Apple apps. They also offered free text messages for taxi companies to reach customers. MusicJuice depends on Google AdSense.
4. Website
Since this is the place that customers visit and decide if they want to do transactions with the company, it must invest significantly in design, making the website attractive and at the same time easy to navigate and take actions. The website performance, reliability, and security are also crucial. With limited cash at the beginning, the website development cost MusicJuice.net almost all of its cash, with only $5,000 left. Since there are millions of websites, it's very hard to design one website that stands out and attracts many people, both the look and feel and the activities. That's why MusicJuice got the feedback of boring website with few activities.

Sunday, February 13, 2011

When Passion for Movies, Technology Collide

The article is from WSJ on Feb 10, 2011 http://online.wsj.com/article/SB10001424052748704422204576130312463129144.html?mod=WSJ_Tech_LEFTTopNews
Netflix definitely pushed several competitors out of business but in this article, Mr. Skorman, an Internet and retail entrepreneur, "believes rental stores with loyal followings in vibrant neighborhoods will survive.", citing that people still need "community element" that online can't offer.
Notice that there are at least two conditions for the brick-and-mortar model to survive: loyalists and vibrant neighborhoods. So, this model can work but only as highly specialized boutiques located in some crowded areas due to the two dramatic limitations that Chris Anderson pointed put in "The Long Tail".


What do you think is the best long-term strategy for Netflix, given the recent growth in popularity and competition in the market for online streaming content?

Netflix is a great example of adjusting its strategies based on the user feedback and changing environments, either internal or external. It started with a pricing model similar to the one used by traditional video stores: $4 per movie plus $2 shipping and handling charge and late fees. Customers voiced their frustrations and Netflix switched to no-late-fee subscription model. A great way to differentiate itself from brick-and-mortar competitors and bring more value to customers. Another example is moving from no direct relationships with major studios to direct revenue-sharing agreements with nearly all of them.
To sustain its position as a giant in the media industry, it needs to pursue a long-term strategy with several moves:
1. Continue the close watch of VOD developments and adjust the strategy accordingly. Partner with key IT players in this trend since as Sarandos explained "What we do is probably 70% science, 30% art." To live up with the three strengths: convenience, value, and selection, Netflix definitely needs to invests continuously in IT.
2. One primary limitation of VOD is content availability. Netflix can add more content to serve the "long tail" community by reaching out to more foreign major studios. As Chris Anderson pointed out in "The Long Tail", with this new economics, in which it matters not where customers are or how many of them are seeking a particular title but only that some of them exist, anywhere. With more diverse content, Netflix can both satisfy current subscribers and attract more new customers.
3. Add more "extras" on Netflix. 
When users select a title to watch instantly, Netflix offers services such as Order Foods and Drinks and Share with Friends. If the users select Order Food and Drink, Netflix can give recommendations based on their history, the statistics of others watching the same title, and the ratings. The idea is similar to movie recommendations. The users can choose to enter their inputs. To make this work, Netflix needs to partner with restaurants. Share with Friends is letting the users' friends know that the users are watching this movie now. The purpose is that their friends can pick that movie to watch, too and later on, they can discuss the movie.
When users select a movie for their queues, Netflix displays the list of their friends who have already watched it. This way, the users can have reviews from their friends, making the reviews more likely to be reliable and valuable. Netflix can partner with Facebook, Twitter, or other social media.
After users watch a movie, Netflix displays All About <the movie title>, for example All About Finding Nemo. It is basically a store where users can buy CDs, books, posters, toys, games, clothes, and anything related to that movie. 
The ultimate goal is to deliver more content instantly and conveniently, more relevant services, and greater viewing experience.  

Tuesday, February 8, 2011


What do you think is the best way for Yelp to monetize the reviews and content they’ve generated, going forward? How scared should Yelp be of Google Hotpot and what should they do to maintain/grow their position?

Yelp should be very concerned with not only Google Hotpot but also other websites such as Yahoo! Local and CitySearch. With many competitors around, Yelp risks being just another review website with not much differentiation. Also, popular websites such as Facebook and Twitter can expand their businesses to include services that are similar or even better than Yelp. Furthermore, these potential competitors have the advantage of well-established names and the network of friends, who users trust much more than several strangers on Yelp.

The first option is charging users for accessing the website. Yelp can charge differently to different groups of users and even give free access to some special groups such as Elite Squad, seniors, or military families. They can also use freemium models. However, due to little switching costs, this will drive away several users, who will likely join competitors’ website. Consequently, in the long term, Yelp is losing market share and can lose ground altogether unless it finds some ways to leverage its website to differentiate itself and bring more values to users.  This may mean to turn this website into a high-end place for quality and trusted reviews by hiring “review experts” or giving more incentives to great reviewers. This option is very hard to implement unless it has very creative, innovative and dedicated technical team. This solution may also work only in the short-term since more and more companies offer free access with similar or even better services.

The second option is to invest more money and effort in the sales team, making them greater in number and more aggressive and creative in reaching out businesses for more advertising money. Different from the first option, this one does not guarantee that the time and money invested for the sales team will be made up and the sales team can even bring in more money. Another uncertainty is that no matter how smart and dedicated the sales team is, they cannot persuade businesses to sponsor if Yelp does not create more incentives for those businesses.

The third option is to partner with other applications such as Facebook, Twitter, and Microsoft Bing. This way, Yelp will get more exposure and build its brand name, thus getting more recognition and users. Business will be more likely to spend advertising money on Yelp. Yelp is already listed as a partner site on Facebook. However, it is hard to get the buy-in from other applications since they themselves can expand their services to include reviews like Yelp. Even though they agree to partner, there can be conflicts of interests or different business strategies that Yelp must depend, compromise or sacrifice. 

The fourth option is to sell the business to whoever is interested. This is probably the easiest way, no longer worries about ways to get more revenues and be competitive. Then, Yelp just gets one-time money and the price offered may be far below what Yelp perceives. This may be a recommendation given Yelp services and the popular business model of free access from several websites. Also, social websites such as Facebook have more advantages since reviews or requests are among circles of friends, who we trust more than strangers on Yelp.

The fifth option is to reach out further, to countries with fewer competitions such as in developing countries, where there are no or few websites exclusive for reviews with the extensive interface and services in compared with Yelp. Yelp can have its website to be translated into local languages and hire local people to do the job. Since Yelp already has the website, it just needs to translate into different languages, so the cost of this translation is insignificant compared with what it had invested in the website. Yelp may use some models similar to “franchise” so it can reduce risks. This option definitely poses some risks. Doing business in a foreign country, especially with different language and culture, requires the management to be very adaptive and Yelp likely to be dependent on the local people to manage and operate.

Among these options, the third and fifth look much better given the harsh competition and this globalization trend. For the best solution, Yelp should implement both of them.

Sunday, February 6, 2011

Some interesting contrasts between Webvan and Yelp

Of course, Webvan and Yelp cases share some similarities, but there are many interesting contrasts partly due to their different industries, both grocery and last-mile delivery industries for Webvan and information goods industry for Yelp. First, during the first few years, Webvan tried to launch its services in several big cities while Yelp just focused on establishing its presence in San Francisco. Second, Webvan spent over $100 million upfront on distribution centers whereas Yelp needed very small infrastructure investments. Third, Webvan needed a complicated supply chain management since it’s in both grocery and last-mile delivery industries. Meanwhile, Yelp was in information goods industry and dedicated to its web content. Finally, Webvan’s direct source of revenues was users and could have made profits if the number of users was big enough. In contrast, Yelp already had a great community of users but could not find ways to monetize this.

Tuesday, January 25, 2011

Why did Webvan fail so spectacularly?

A spectacular case study!
For the first part of the question, here are some major reasons why Webvan failed.

From the supply chain management perspective, we need to consider the six performance drivers, namely facilities, inventory, transportation, information, sourcing, and pricing. Among these, Webvan’s costs of facilities, inventory, transportation, and information (including software) are much higher in comparison with traditional supermarket supply chains. For sourcing, Webvan needed its employees to pick items for orders instead of customers doing this at a bricks-and-mortar store. So, it added extra labor costs for handling customer orders. All of these higher or extra costs were applied to the grocery industry, where margins were only 1% to 1.5%. To make the matters worse, Webvan advertised that its prices were 5% lower than conventional stores. All of these resulted from its hope that the number of customer accounts would be high enough to make profits after three or four quarters. In reality, the number was far below the forecasts and the company kept losing money. Clearly, Webvan’s supply chain design was too expensive to be profitable and too elaborate to operate efficiently and effectively.

From statistics and forecasting perspective, Webvan came out during the heyday of Internet companies when there were not enough stories of failures from history to tell and learn from. Thus, timing also played a role here with overly optimistic numbers and forecasts such as 5% of US households would buy groceries online in a few years and online grocery market would be worth $3.5 billion in 2000 and $6.5 billion by 2003. In this high spirit, Shaheen saw the market as $1.5 trillion, an IDC projection for 2003, which encompassed all web-based purchases. Based on these fantastic numbers, Webvan CFO insisted that Webvan would be “highly cash generative” and that the DCs were likely to operate at breakeven capacity within five quarters of being launched. In reality, it had not hit this target after six quarters since the launch.

From strategy perspective, the management team was too confident and ambitious. They wanted to do everything everywhere in a huge scale. Consequently, they went against their original strategy of providing a more cost-effective solution. They acted hastily in building huge, expensive, and complicated DCs. At the same time, they invested money for plans to expand into various US regions at the same time. They also announced projections that were almost impossible to be realized such that if everything went according plan, Oakland DC would be profitable within 6 to 12 months and other DCs might break even in 60 days. Even a tiny company never has everything going according plan, let alone company with complicated information systems and huge infrastructures as Webvan. Naturally, they should have been prepared to get several unexpected problems and thus, that statement should have never been made. They hoped to get 8,000 orders a day from Bay Area DC to make operating margin target of 10% to 12%. In reality, after six quarters, the averaged number of orders was only 2,160, too far below the projection.

Note that all the changes later, including partnerships and Homegrocer acquisition, could not save Webvan due to either ineffective and inefficient designs and implementations or being done too little too late.  

And the second part,  here are some major reasons why Webvan failed spectacularly.

First, its funding happened so fast and spectacularly. In 1999, it was the most funded for an Internet company with $400 million. Its first day of trading, at one point, it mounted to $15 billion capitalization. It raised a total $800 million. All of this funding and market valuation happened for a company with only $4 million in revenue at that time.

Second, its dream team was included so many senior executives experienced in a broad range of industries with well-established companies such as Borders Books, Goldman Sachs, Oracle, and FedEx. With these diverse experts, who could imagine Webvan would fail?

Third, for the three straight quarters in 2000, it had been voted the best online grocer of 12 in a survey. This means, customer satisfaction was achieved well.

Fourth, in the last quarter of 2000, only 6 months before it closed for good, it posted a gross margin of 27%, highly competitive with large conventional grocers.

Finally, the time it took to fail was also dramatic. It closed its doors less than two years since its heyday on November 5, 1999.