By: Steve Jurvetson and Tim Draper
Jan. 1, 1997
"The award for Internet marketing buzzword of the year goes to 'viral marketing." - Iconocast, December 16, 1998
A lot of the energy behind the Internet is the ability for everyone to be a publisher. Consequently, we are in a land grab for precious spectrum - people's attention. Attention is finite. Rising above the noise of a thousand voices requires creativity. Shouting is not very creative. Just hanging up a web shingle and hoping for visitors is not very creative. Rather, new companies can structure their businesses in a way that allows them to grow like a virus and lock out the existing bricks and mortar competitors through innovative pricing and exploitation of these competitors' legacy distribution channel conflict.
In 1996, Sabeer Bhatia and Jack Smith pioneered a great new product category -- free web-based email. But many great ideas and great products have withered on the vine. The special catalyst for Hotmail's torrid growth is what we at Draper Fisher Jurvetson have come to call "Viral Marketing" -- not because any traditional viruses are involved, but because of the pattern of rapid adoption through word-of-mouth networks. Viral Marketing powerfully compounds the benefits of a first-mover advantage. And it's something we eagerly look for when evaluating any Internet startup company. As a founding investor in Hotmail and a member of their board of directors, we think Hotmail is a great case study on the impact of the Viral Marketing strategy over its full life cycle.
Hotmail's Amazing Growth:
Hotmail grew a subscriber base more rapidly than any company in the history of the world ...faster than any new online, Internet, or print publication ever.
Hotmail is the largest email provider in the world.
In its first 1.5 years, Hotmail signed up over 12 million subscribers.
A traditional print publication would hope to reach a total of 100,000 subscribers within a few years of launch. Hotmail signs up more than 150,000 subscribers every day, seven days a week.
Every Hotmail subscriber, without exception, has filled out a detailed demographic and psychographic profile including occupation and salary. This is an unprecedented supply of personal information.
Yet, from company launch to 12 million users, Hotmail spent less than $500K on marketing, advertising and promotion. This compares to over $20 million spent on advertising and brand promotion by Juno, Hotmail's closest competitor with a fraction of the users.
Other companies may have distributed more unit volume of product than Hotmail did in their first year - especially when releasing upgrades or brand extensions to an established franchise. But for a new entrant with a new product, the challenge is more daunting. "Subscriptions" have their own challenges as well. Users face a trust decision in deciding to share their private information and email with an online entity. And the user may not be certain that the end product is worth the effort. These are barriers to adoption in the subscription model. How did Hotmail overcome these barriers as an undercapitalized startup? Viral Marketing.
Hotmail originally approached us as JavaSoft, Inc., a web database tools company, and, as Business Week recounted: Sabeer and Jack went to see "Draper Fisher Jurvetson, but the investor was unimpressed by their idea for database software for the Net. As they were packing up to leave, [the VCs] asked: 'Do you have any other ideas?' Sabeer said they'd noodled over a scheme to offer free, advertising-supported E-mail over the Web. A week and a half later, the venture capitalists ponied up $300,000, and Hotmail was born." (BW, August 25, 1997)
In our next meeting, Tim Draper suggested that they should append an advertising message to every outbound email: "P.S. Get your free email at Hotmail. It was very contentious at the time. Would users balk at having this automatic addition to the content of their private messages? Hotmail tempered the idea by clearly demarcating the promotional plug, and removing the "P.S." Nevertheless, every outbound message still conveyed an advertisement and a subtle implied endorsement by the sender - the recipient knew that the sender was a Hotmail user, and that this new free email thing seemed to work for them. Each new user becomes a company salesperson, and the message spreads organically.
Hotmail's business model maps well to the medium. By contrast, Juno does not map well to the medium, and they have already spent $20 million in advertising. Hotmail did not spend the money, yet gained over three times as many users in half the time.
Elements of Viral Marketing:
The Hotmail adoption pattern is that of a virus - with spatial and network locality. People typically send e-mails to their associates and friends; many of them are geographically close, and others are scattered around with clusters in areas of high Internet connectivity. We would notice the first user from a university town or from India, and then the number of subscribers from that region would rapidly proliferate. The beauty of it is that none of this required any marketing dollars. Customers do the selling.
Digital viruses can spread internationally more rapidly than biological viruses that rely on the physical proximity of hosts for their spread - via a sneeze or handshake. Hotmail is the largest email provider in Sweden and India despite the fact that they have done no marketing of any sort in these countries. It's a happy day when you discover your business has displaced several entrenched competitors to become the market share leader in a country you have never visited. What's more, Hotmail is used in over 220 countries, despite the limitation that it is only available in English.
Viral Marketing captures the essence of multi-level-marketing and applies it to all customers - the "word-of-mouth" spread of the Hotmail message is involuntary. And it's more powerful than many other marketing techniques that lack the implied endorsement from a friend. Hotmail had "Free Email" buttons on several other highly-trafficked web sites, but they generated comparatively negligible numbers of subscriptions. Juno has shown that advertising is relatively cost-ineffective. It is hard to spend your way to Hotmail-like growth. The snowball effect is a mechanism to greatly leverage a first-mover advantage. Whenever a product involves people other than the purchaser, then there is an opportunity to market to potential new customers. It is no surprise that Amazon encourages its customers to send a book as a gift to a friend. When the recipient receives the gift book, the packaging contains a flyer for the amazon.com service. Similarly, whenever someone uses iShip.com to send a package, the recipient will learn about how iShip.com can save them money on their shipping needs.
As more Internet and Intranet applications move beyond computation to embrace communication, the Viral Marketing strategy has wide applicability. E-commerce, groupware, community, messaging and promotions businesses can all use these techniques to further the Internet explosion.
In addition to Hotmail, we have seen a similar viral communications approach taken by Mirabilis, an Israeli-based company that signed up twelve million instant messaging subscribers to its ICQ service before its acquisition by AOL for almost $300 million. To use ICQ, both parties need to download the client software. So each subscriber ends up emailing their friends to solicit them to engage in this new communication channel. The friends in turn, experience the product first hand, and may repeat the pattern. It's like the old shampoo commercial refrain "they told two friends who told two friends, and so on." While powerful, these dyadic communication products recruit new customers one by one.
A sneeze releases two million aerosol particles. In the digital domain, this can get very interesting. For example, Tumbleweed Software enables secure e-mail delivery of documents or newsletters to a huge numbers of recipients. Every recipient also gets a web link to the enabling Tumbleweed service. So when a single new customer starts to use Tumbleweed, thousands of potential new customers receive the Tumbleweed pitch.
This viral broadcast model can be creatively applied to a variety of products, such as web-hosted address books, calendars, list servers and news group readers. Homestead enables families, sports teams, alumni organizations, and other affinity groups to create rich and private "community" web sites. When someone builds a Homestead, the system facilitates them to broadcast an invitation to the members they would like to recruit to participate in their Homestead. A new recruit may belong to several affinity groups, and may find the service compelling enough to replicate in these other circles - thus proliferating the Homestead platform.
The power of this approach has been demonstrated in the junk email domain. Have you ever gotten one of those email chain letters that urge you to forward it to as many people as possible? Often shrouded with a bogus virus warning or a charitable cause, these messages rapidly spread throughout the globe until people have received multiple copies.
These junk e-mails are like digital graffiti in that the people that create them want their "tag" or message to be seen by as many people as possible. Traditional graffiti "artists" choose targets like trains and buses to maximize their exposure. Similarly, many computer virus authors are seeking to promote their name, and they seek maximum exposure - on the PC. You don't see many viruses on niche computer platforms (as the Macintosh market share has dwindled, so too have the number of new Mac viruses). This personal quest for fame, while annoying, is not too different from the desire of many businesses for brand awareness.
New businesses also want to let the world know their name, but they do not want to be annoying, and thus, they must provide a compelling reason for customers to help spread the word. In all of these viral businesses, the balance between the customer value provided by the network and the concerns surrounding privacy will be a delicate one.
For many network applications - from ICQ to the traditional fax machine - the value of the network, and the value that each member realizes increases disproportionately as more people join the network. The first fax machine customers were delighted to see more people buy compatible machines. A company that can provide a strong enough incentive for customers to share their massive lists of personal contacts - whether for communications or community - will have a powerful viral opportunity at their disposal. A good virus will look for prolific hosts (such as students) and tie to their high-frequency social interactions (such as email and messaging). Viral Marketing is most powerful when it taps into the breadth of its customers' weak connections to others. Tapping the customer's entire address book is more valuable than just reaching their best friend.
Viral Marketing Strategies:
The typical viral entry strategy is to minimize the friction of market entry and proliferation with an eye to building in hooks and barriers to switching for customers. If the service is trying to blatantly monetize its subscriber base in every way imaginable, new users will be reluctant to spread the word. Therefore, many of these services are free and light on the revenue generation in the early days of their rapid proliferation. When we first invested in Four11 and Hotmail, we could not say with certainty how they would ultimately monetize their subscribers. We brainstormed several possible scenarios for how they might eventually exploit their large audience and market position as a communications hub. But in the viral growth phase, the simple banner ad seemed the most innocuous.
In an extreme example, prior to their acquisition by America OnLine, ICQ's CEO took delight in the fact that they not only had no revenue, but had no current plan for revenue. This is not to say that businesses without revenue prospects are necessarily attractive - just that people's attention (or "eyeballs") have proven to be monetizable in every media.
A company that can choose to delay revenue maximization (e.g., by not burdening their service's clarity of purpose and speed of download with excessive ads and promotions) may find that they can exploit a first mover advantage in the Internet land grab to gain a dominant market position. This is one of the reasons so much VC money flows into these Internet start-ups.
The Internet is a wonderful substrate or petri dish for the proliferation and replication of intellectual property. A good idea can spread more quickly over the Internet than had ever been possible before in the physical world, where manufacturing and distribution fundamentally limit the rate of product adoption. Especially in the Internet era, a company's competitiveness seems to depend on its velocity of thought and action. Companies can grow more rapidly than ever before, but so too may they suddenly die from obsolescence. The critical differentiator is whether the company has built in switching barriers for its customers and barriers to entry for its competitors. Rapid growth is of no value without customer retention.
Whenever we consider an investment in an Internet startup company, we strategize about customer switching barriers, and the impact of the inevitable arrival of competitive imitators. We have witnessed a particular entrepreneur that has used an offshore development team to quickly copy one of our portfolio companies' lead into a new market - three times in a row! First they copied Four11's directory services, then Hotmail's free email, and now Kana Communications' wildly successful email-based customer support software. The Internet supports an ecology of organisms, and the "fast follower" is a classic form.
The ability to rapidly recruit subscribers creates market value - but only if a company can retain these customers over multiple visits to their site. In the public markets, many of these consumer Internet companies are being valued at $20 to $100 per subscriber. Investors and bankers are approximating the lifetime economic value of these subscribers discounted back to the present day. How many ad dollars will they generate over lifetime? How much will they buy? A huge variable in this economic equation will be the customer retention rate.
Are you like a subway station with banner ads flying by the commuters who are just trying to get to their destination, or are you like a cafe where customers mingle and feel like they belong? Web communities, e-mail, personalization, contact lists, calendars, personal web pages - these are all "sticky" applications that help retain customers. This is where most of Yahoo and the other portals' energy has gone in the last two years. That's why Yahoo bought Four11 for its RocketMail service and Microsoft bought Hotmail.
An interesting side effect of geometric growth is that by the time a virus spreads to the point of being an epidemic, its growth curve relative to a new entrant is somewhat daunting. Hotmail was doubling in size each month, but it took several months to reach one million users. Until then, they were under the radar screen of many potential competitors and acquirers. By the time the industry came to realize that free web-based email was indeed a hot idea, Hotmail was adding one million new subscribers per month, and that growth rate was accelerating. A new fast follower would start small and have to grow for several months to reach one million subscribers. But in that same time, Hotmail would have grown to 10 million subscribers. So although Hotmail's followers grew geometrically as well, the absolute difference in subscriber bases widened every month (while the ratio remained approximately constant).
Absolute size matters. One significant effect of Hotmail's absolute size is that their efficiencies of scale allowed them to be the lowest-cost e-mail provider on the planet. Server utilization and bandwidth pricing improved with growth. Also, the perceived gorilla in a category tends to get the dominant share of the business and financial partnerships. Many advertisers and media companies do not want to spend time with sub-scale properties. All of this makes it tougher for the smaller new entrant. It also skews the make vs. buy decision toward "buy" for the large portal companies, which realized en masse that they wanted an email solution with proven scalability. A similar buying frenzy is currently underway for "community-building" web sites, such as Tripod, PlanetAll and Homestead.
A challenge for the hyper-growth gorilla is scalability. On a technology level, server scalability is a critical concern. Fortunately, companies like Hotmail are turning software into a service. What was sold as email servers and clients is now offered as a web-based service where the customer need only have a standard web browser. This makes product upgrades a lot easier; Hotmail can upgrade its server software several times a month without involving or in many cases, even notifying its large customer base. The customer still uses the same browser.
But once one problem is solved, hyper-growth tends to uncover new scalability bottlenecks. Often the young Internet company finds that its growth is constrained by its ability to hire good people. This is why many of these companies try to engineer around people-intensive elements of their business.
New Distribution Channels:
New companies are often the primary beneficiaries of new distribution channels, as Dell has shown in the PC industry. Dell just passed Compaq to become #1 in desktop PC shipments. Compaq's legacy channel partners prevent it from entering the lower-cost mail order channel. For many businesses, the Internet is a still lower cost channel of distribution.
By lowering prices or offering free products, and employing a "market shrink" strategy, the new entrant can make it very painful for established companies with established distribution relationships to follow them. Although the new market size may be smaller, driven by Internet price efficiencies, the new entrant can gain significant share by restructuring the basis of competition. There may be less revenue in a free email market, but it's tough for Eudora and companies based on selling client software to follow Hotmail's lead.
Viral Marketing provides a new distribution channel for almost any Internet application. Although it naturally lends itself to free "communications" or network applications, Viral Marketing could also be applied to traditional stand-alone software to accelerate the word-of-mouth spread of good software. How might this work? Release Software can embed in just about any software application an e-commerce engine for electronic software distribution and "try before you buy" purchasing. When a customer gives the application to a friend, it triggers the embedded sales agent to offer a 30-day trial period, after which the new user has to pay for the software. It turns software piracy into a sales opportunity.
Release can also credit multiple distribution partners involved in the sale with a percentage of the transaction. To engage viral marketing, the customer would be treated like a distribution partner. So, in the ultimate pyramid scheme, if a software company were willing to pay a two percent sales commission, a customer could be credited 1% of the sales price for a copy she gives to friend when they buy, and 0.5% for her friend's friends, and 0.25% for three levels down, and so on - thereby spending no more than 2% of sales on "word-of-mouth" promotions. Netcentives, a web currency innovator, could be added to this equation to allow users to accumulate frequent flyer miles for spreading a software application to colleagues and friends.
This could lead to some interesting consumer behavior. Customers would have an incentive to post software to their favorite download sites or other distribution outlets to maximize the total sales of their copy of the application. Not only would the customers help resell product directly, they would innovate and discover new distribution networks. Perhaps these should be called "self-organizing viral distribution networks." Established companies are unlikely to experiment with them, because of channel conflict with their legacy partners.
Where might this all lead? We are still looking for the emergent intelligence of the hive. An ant colony exhibits a higher order of intelligence than that of its individual members. We don't look at a neuron and think of it as being very smart. In many ways, we are the neurons on the Net, and the network applications that take advantage of that collective intelligence have not been developed yet. The Santa Fe Institute argues that computer viruses are a form of artificial life. Perhaps viral marketing can also find an evolutionary form.
But in the meantime, the Hotmail juggernaut just keeps on growing - quietly, and consistently on its own momentum. Hotmail now has over 30 million subscribers. As a technological dislocation, we believe that the Internet provides an unfair competitive advantage to nimble startup companies. A good idea can spread like wildfire if its business model maps to the medium. Viral Marketing adds fuel to the fire.