Lessons from Korea
The following article written by Brendon Kim, one of my partners at Altos Ventures, was published by the Software Development Forum in August. I thought I'd post it here also.
Korean Petri Dish
Take 16 million households composed of 47.5 million people and give 75% of them broadband Internet access – access that is at times 20X faster and much cheaper than what we find in the United States – and interesting things will happen. South Korea has often been described as a Petri dish for developments in the Internet.
Over the past few years, the partners at Altos Ventures have spent a great deal of time thinking about, visiting and building relationships in Korea. We have observed that, indeed, interesting things are happening. Many of the most successful Internet companies in Korea have pioneered new behaviors and new business models and while not always directly applicable, there are lessons for Silicon Valley.
We have also observed that most of the better Korean startups share some common characteristics – characteristics we like to see in U.S. companies. NHN Corp. is testament to the notion that local knowledge will beat global expertise. NHN runs Naver.com the leading search engine and portal in Korea. Google and Yahoo may dominate in the United States, but they are mere also-rans in Korea. Naver relies on the familiar advertising revenue model, posting $228 million in online ads last year, almost 40% of the market. What are not so familiar are the results of a search on Naver. In addition to the standard list of links to websites, Naver posts links to other types of documents such as blogs, maps, dictionary entries and books. All well and good, but the most distinctive feature and arguably the reason why Naver is more popular than any other search engine is Naver’s pioneering use of user-generated content – truly local knowledge. Naver encourages users to post answers to queries from other users. Responses are rated by other users and catalogued. Responders are also ranked and given stature in the community. Ask a question about anything, say raising a puppy, and often, the most complete and popular answer comes from another user. Other engines have started to offer this feature, but Naver has such a commanding lead in the number of user entries, that it is difficult, if not impossible, for others to catch up.
Years before young Internet users in the United States were creating profiles in Myspace, Koreans in their teens and twenties were posting homepages in Cyworld. By some estimates one out of four Koreans or 90% of all Koreans in their twenties have a Cyworld homepage. When meeting for the first time, it is not uncommon for people to exchange Cyworld addresses rather than email addresses or phone numbers. The community has developed its own culture and customs. For example, ignoring comments from visitors to your home page is considered extremely rude and you can be pilloried for it. This encourages people to be very active participants in the community.
Cyworld’s revenue model is also pioneering. Basic services are free, but the site generated close to $200 million in revenues in 2005. A large portion of revenues comes from the sale of dotori (acorns), Cyworld’s virtual currency. Users spend dotori to buy virtual objects to decorate their homepage “room” and accessorize their avatars, one upping each other. Rooms might be enhanced with a digital TV or couch or flowers. Avatars might be dressed in the latest fashions or jewelry – all for a few dotori, which eventually add up to a lot of real money. Members are known to spend hundreds of dollars a month on these digital items and countless stories have been told of teens that have been punished by their parents for going too far.
The items selling business model generates revenues (and profits) for the Korean online gaming company, Nexon. Nexon posted well over $200 million selling digital avatars and accessories last year. Among other games, Nexon developed Kart Rider, an online hit in which players race each other in cartoon graphics gocarts. The company claims that nearly one out of four Koreans have played the game and that over 200,000 players are racing at any given time. Races have even been broadcast on cable. With the support of real world sponsors, some have made Kart riding their profession. Nexon does not rely on any subscription revenues; Kart Rider is free to play. Instead, the game charges you for items – accessories, such as missiles to take your opponent out of a race, balloons that lift your kart out of a missile’s way, smoke bombs that cloud your opponents’ vision, goggles that help you see through smoke and performance enhancements to make your cart perform better. The game is simple to play and just minutes a race, so it is easy to get hooked. Every time you lose to a better equipped player, the more you want to level the field or gain advantage and the more you want to reach into your wallet.
Pandora.tv is Korea’s leading video destination on the Web. (Pandora is an Altos Ventures portfolio company.) The company was established months before YouTube in the US. Early on, Pandora’s founder, Peter Kim, saw that some of the most popular sites in Korea revolve around self expression. Cyworld was an obvious case in point and to some extent, so was Naver. So, unlike many other video sites, Pandora was designed to be more about self expression and less about video sharing (although you can certainly do that too). Pandora is mainly organized around channels that members make their own with a collection of videos that expresses their characters or interests. The site posted 300 million page views in June, doubling page views in just three months.
Pandora focused on generating revenues from the start and experimented early with several business models. Pandora derives revenues from a number of sources: pre-roll advertising, where an ad is played before a requested video; consumer subscriptions, where users pay for higher bandwidth; content distribution fees; and “brand channel” sales, where businesses buy a channel for commercial use. Brand channel customers include major corporations such as Reuters, small businesses, churches, schools and even politicians. At the time of our investment, Pandora was already at cashflow breakeven with minimal outside funding.
NHN, Cyworld, Nexon and Pandora innovated in varied ways but they shared some similar characteristics when first starting up. Based on our observations, many of the better Korean companies also share these qualities. At Altos we look for these same characteristics in U.S. companies.
- The company is capital efficient. The venture industry in Korea is not as well developed as that in the US and early stage venture capital is scarce. As a result, most entrepreneurs are forced to bootstrap, looking for revenues early and keeping costs low. Capital efficiency is built into the DNA of the company and the team.
- The company continually experiments and tinkers with the business. Koreans are said to have a cultural bias towards action and learning by experience over data collection. Companies innovate over time by experimenting, learning and implementing the learning.
- The entrepreneur is extremely focused and single-minded, much like a hedgehog. For a country of 47.5 million people, the startup community is surprisingly small and reputation is everything. The entrepreneurs we like have a reputation for being thoroughly committed, headsdown, execution-oriented and interested in creating an enduring business with a solid business model.
In the last several years, Korea has emerged as an innovator in semiconductor manufacturing, wireless, consumer electronics and the Internet. The country is not about to stop here. Korea is in the midst of a major five year government sponsored technology program, the 8-3-9 Initiative, designed to keep Korea on the cutting edge of technologies such as IPv6, 3G/4G wireless, wireless broadband, mobile broadcasting, RFID and digital television. Korea is enjoying its recent role as innovator and Petri dish and we would benefit by keeping an eye on the growing developments there.
