jump to navigation

Three use cases for virtual goods January 28, 2008

Posted by jeremyliew in communication, mmorpg, self espression, social games, social gaming, social media, social networks, virtual goods, virtual worlds.
16 comments

Last week I estimated that Facebook is doing up to $15m in revenue per year from its digital gifts business. I noted that there are three use cases for digital goods. I thought that it might be useful to go into each of those use cases in more detail.

Digital Gifts

Facebook, HotorNot. Live Journal and Dogster are all examples of companies that have rolled out virtual gifts that members can send each other.

Virtual gifts are most effective in the context of communications, especially high volume communications environments where it can be difficult to get attention. By paying real money for a virtual gift, the sender of a message signals that they are more eager than most others to be heard. As James Hong has noted:

The utility gained from this is one of SIGNALING. What is it you are trying to signal? In the case of HOTorNOT’s virtual flowers, one is trying to signal extraordinary levels of interest. A user on the site can say “yes i’m interested” to every other person on the site because it costs nothing (but time) to click “yes” on people’s profiles. However, it is presumed that money IS a limited resource. By spending money in order to purchase a flower, becaues the # of flowers I can afford is finite, it signals to the recipient that S/he is very very EXTRA special. So we chose to price the flowers high…

So basically, because they’re so expensive and less people are willing to send them is what makes someone who RECEIVES them DIFFERENTIATED. The flowers have REAL value to the receipient [sic], and therefore real value to the sender who is gonna get props for sending them…

We found, last time we ran the numbers, that sending flowers increased the likelihood of a “double match” on our system by 4x.. meaning as a signal, they are well received and really work.

Dana Boyd also points out that gifting opens up the opportunity for reciprocation:

Gifts are part of status play. As such, there are critical elements about gift giving that must be taken into consideration. For example, it’s critical to know who gifted who first. You need to know this because it showcases consideration. Look closely at comments on MySpace and you’ll see that timing matters; there’s no timing on Facebook so you can’t see who gifted who first and who reciprocated. Upon receipt of a gift, one is often required to reciprocate. To handle being second, people up the ante in reciprocating. The second person gives something that is worth more than the first. This requires having the ability to offer more; offering two of something isn’t really the right answer – you want to offer something of more value.

This communications context for gifting underscores our finding that holiday themed facebook gifts sold 5x better than average facebook gifts. Holiday gifts come with a built in message. Getting a Santa Hat as a virtual gift is much easier to understand than getting a beach ball as a virtual gift.

Self Expression

Social network users love to personalize their profile pages, whether they be MySpace users, Bebo Users, Orkut users, yes, even Facebook users! Many web sites and virtual worlds have found that users are willing to pay to personalize their web representations, as evidenced by Gaia, Habbo Hotel, Meez, CyWorld, Second Life, Tencent and many more. These are businesses that in some cases are making tens, even hundreds of millions of dollars in revenue by selling virtual goods to personalize virtual avatars, apartments/hompies and the like. As Fred Stutzman notes:

In the SL and Cyworld model, the motivations are built on very sound logic. People like to buy stuff for themselves that makes them look cool. Since online identity is primarily about the representation of self, people will pay to differentiate themselves.

This is not so different from the real world, as evidenced by the continued growth of the luxury goods industries (including apparel, jewelry, luxury cars, watches etc).

Increased Functionality

The third common use case for digital goods is for users to buy increased functionality or power. This is ften within the context of a game. As Susan Wu noted on Techcrunch:

Each day, thousands of transactions take place via markets such as eBay for virtual swords, currency, or clothing across a multitude of virtual world environments. For people who purchase virtual items such as swords or armor, buying these items increases the overall satisfaction she receives from spending time in this virtual world / online community / online game. For example, struggling along as a level 20 character might give her 20 units of personal satisfaction per hour, whereas progressing as a level 20 character with a very powerful sword could confer 50 units per hour. In this case, she would be willing to pay the equivalent of whatever amount generates an incremental 30 units of personal satisfaction for the sword.

I’m an avid player of multiplayer online games. A couple of years ago, I spent 10 real dollars to buy 1 million gold in a game [yes, it was legal and part of a world where real money trade is not prohibited.] My friends mocked me and told me I was throwing money away, so I tried to explain it to them: 1 million gold would give me 20 hours of entertainment. If I were to go to the movies, 10 real dollars would buy me 2 hours of entertainment. Assuming that 1 hour of movie watching entertainment gives me the same personal satisfaction as 2 hours of game playing enjoyment, I would have been willing to pay $50 in exchange for that 1 million of virtual currency. In fact, I felt like I had gotten a bargain paying only $10!

Buying gold, or RMT as it is known in the gaming world, exists for just about all MMOGs, whether sanctioned by the game or not, and has spawned businesses such as IGE and Sparter.

Maintaining Scarcity

One key to the success of digital goods business models is to maintain the scarcity of the digital goods. Since digital goods are digital, they cost nothing to copy. Free copies of digital goods would reduce demand for paying for the same item. In a closed system, it is easier to maintain scarcity. The company controls the supply of all digital goods completely.

In an open system, the situation gets more complex. If users (or app developers) can create content that can be injected into the system (whether it be a website, a profile page, a virtual world or whatever) then this can readily blur the lines for self expression and virtual gift type digital goods. If it is easy, or even possible, for users to mimic digital goods, and if this creates confusion or uncertainty about which goods are “premium”, then digital goods can become devalued. To some extent, this has happened in Facebook, where free gifts applications have proliferated. This has probably dampened the sales of Facebook’s digital goods.

Myspace has never seen a digital goods model take off because the completely open profiles make it impossible to differentiate a paid digital good from a free copy. Both end up being a .swf file or an image, with no ability to differentiate. This may be one of the reasons that CareBadges (a MySpace widget bought by donating to a cause) had difficulty achieving real scale in MySpace.

Even in a closed system, if there is uncertainty about whether a digital good is free or not, this can inhibit the sale of digital goods. Demarcation lines between free and premium need to be clear, consistent, and obvious to even a new user to sustain the value of premium digital goods.

Increased functionality is one area where this demarcation is easy to maintain because this is typically an area that is within complete control of the website or virtual world.

As systems grow, the number of digital goods in circulation can rapidly grow as well. This can slow the demand for further digital goods purchases as longer tenured users decide that they have “enough” self expression or increased functionality. Good system designers will create “sinks” for digital goods to maintain a continued demand for digital goods from even long tenure users. These can include wear and tear on digital goods (Cyworld furniture gets worn as it is sold, HotOrNot flowers die over time), or risks of virtual goods being destroyed when they are upgraded (ZT Online).

I’d be interested to hear from social media and social game designers building digital goods business models to see if they have comments on this taxonomy or examples of successes and failures of digital goods models.

Lightweight self expression for the general public November 21, 2007

Posted by jeremyliew in blogging, communication, Consumer internet, product management, self espression.
2 comments

MIT Technology Review has two good articles about microblogging in the November/December issue. (Both are behind a free registration wall.) The puff piece on Evan Williams and Twitter notes some of his thoughts on micbroblogging:

The criticism doesn’t seem to bother Williams, in part because he’s heard it before. “Actually, listening to people talk about Twitter over the last few months, you hear that almost all the arguments against it are the exact same arguments that people had against Blogger,” he says. “‘Why would anyone want to do this?’ ‘It’s pointless.’ ‘It’s trivial.’ ‘It’s self-aggrandizing bullshit.’ ‘It’s not technically interesting.’ ‘There’s nothing to it.’ ‘How is this different from X, Y, and Z that’s existed for the past 10 years?'” Indeed, there were blogging tools available when Blogger was released, and others have emerged since–including TypePad from Six Apart, which offers more features. But none has the simple appeal of Blogger, and none is as easy to use. These were the reasons Blogger was such an important force in the blogging revolution.

There is an interesting idea at the heart of all this, and that is the idea of innovation through removing features. By focusing on a subset of core functionality, both Blogger and Twitter (and the other microblogging startups, as well as Facebook’s status) have made the user interaction much lighter weight. In my experience at AOL, Netscape and IAC, lightweight interactions generally work better with the general public.

Last year Gartner predicted that blogging would peak in 2007:

The analysts said that during the middle of next year the number of blogs will level out at about 100 million. The firm has said that 200 million people have already stopped writing their blogs… Gartner analyst Daryl Plummer said the reason for the levelling off in blogging was due to the fact that most people who would ever start a web blog had already done so. He said those who loved blogging were committed to keeping it up, while others had become bored and moved on.
“A lot of people have been in and out of this thing,” Mr Plummer said. “Everyone thinks they have something to say, until they’re put on stage and asked to say it.”

Microblogging removes some of the pressure to write substantive posts, making it a lighter interaction that is easier to keep up.

The public’s preference for lightweight self expression is part of what has made widget providers (such as Rock You, a Lightspeed company), profile layout sites (such as Free Code Source) and quiz sites (such as Quizilla) so successful.

Business models for apps and widgets November 16, 2007

Posted by jeremyliew in ad networks, advertising, business models, facebook, myspace, open social, platforms, self espression, social media, social networks, user generated content, widgets.
9 comments

This afternoon I spoke to the Stanford class on Creating Engaging Facebook Apps.

As I said at Web 2.0 expo, building big businesses online is hard work. While it isn’t hard to start an app company, especially as a single developer ($250k in revenue) or even to support a small team ($2.5m in revenue), it gets quite hard to scale revenues to $25m/yr.

Assuming a 5% daily active rate and 3 pageviews per visit, an app developer with a $0.50 RPM would need to get to 926m installs to get to $25m run rate. Compare that to the app with the most installs on Facebook – Slide’s Superwall which has around 20-21m installs. Clearly, broad reach app developers need to develop (i) multiple (ii) high engagement apps [ie higher active rates and pageviews/visit than these assumptions] (iii) across multiple social networks to be able to get close to this revenue target. (RPMs will likely be higher for companies with a direct sales force as well, so the target isn’t quite as high, but you get the point).

Under the same activity and pageview assumptions, an app developer with a $10 RPM would need 46m installs to get to $25m in revenue. Apps with endemic advertising opportunities can easily realize this level of RPM but will still need to be in multiple social networks to get to those levels of installs. It doesn’t make sense to limit your world to being a Facebook app. Social network platforms are avenues for distribution, and app developers should be taking advantage of all of them.

One of Lightspeed‘s portfolio companies, Rockyou, is taking the former approach. Another, Flixster, is taking the latter. Both seem to be working so far.

I also did a similar analysis for digital goods business models in the presentation. Here is a link: Stanford Facebook Class presentation

Facebook for engagement; Myspace for self expression August 28, 2007

Posted by jeremyliew in business models, engagement, facebook, myspace, self espression, social media, social networks, widgets.
4 comments

As Techcrunch, Mashable, Venturebeat and others have noted, Facebook is preannouncing a number of changes to its APIs, including a shift to user engagement as the way it presents apps in the directory:

This week you’ll see us shift our application directory metrics to a focus on user engagement. This will help inform users as they make decisions on which applications to add as well as shift developer focus to engagement rather than total users. More specifics will be available as we roll out these changes this coming week.

The focus on engagement is a reflection of how app developers are already behaving today, especially when compared to the widgets being built for MySpace.

Both Facebook and Myspace disallow advertising in the widget/apps appearing on profile pages. But Facebook allows application developers to control the canvas page and place ads on those pages, while giving access via the APIs to the social map. As a result, there are already quite a number of companies reputed to be doing over 100m “pageviews” (canvas views) per month, including iLike, Flixster, Rockyou, Slide, Texas Holdem, HotOrNot, and others. [disclaimer – Rockyou and Flixster are Lightspeed portfolio companies]. The companies with a lot of installed facebook apps are all already pursuing engagement, even before Facebook’s change in the application directory. It’s in their business interest to do so.

In comparison, Myspace is still primarily about self expression. Click through rates from widgets on Myspace are dramatically lower than on Facebook apps. And more importantly, when a user clicks through on a Mypsace widget, they stay for less pageviews than a Facebook app. Because the interaction takes place off site (off Myspace’s site) and because there is no access to the social map, the primary activity is for a new user to create their own widget. There is limited ability for a user to interact with an existing widget because there is such a low level of shared data between the widget and Myspace.

Facebook has only a third the pageviews and UU of Myspace in the US. All else equal, you would expect all of the the top Myspace add-on sites according to Mashable to have over 100m PV/month if Facebook has already been able to generate so many at that scale:

2006 Top Myspace add-on sites

But according to Comscore, only half of the these ten sites have more than 100m pageviews/month, and in at least some cases (e.g. Youtube), the primary traffic driver is not MySpace.

PVs for myspace addons

Given the striking difference in engagement levels, its not surprising that all the other social networks are considering a platform strategy of their own.

Facebook apps are providing new stages for “performance” by users July 17, 2007

Posted by jeremyliew in communication, Consumer internet, facebook, Internet, performance, self espression, social media, social networks, user generated content, web 2.0, widgets.
10 comments

Its now widely agreed that the two most common behaviors on social networks are self expression and communication.

Most of the online revolutions have been driven by new forms of communication. This started with Usenet and BBSs back before there was an internet, moved through the chat rooms of early AOL, the mainstreaming of email and the instant messaging revolution with AIM and ICQ. Communication has always been a large portion of overall time spent online because it drives both frequency of visit (people check for communications often) and depth of visit (reading and responding to your messages takes time).

Social networking is no exception, and that is what has driven the extraordinary pagesviews for the top social networks. In the case of social networks, the primary communications channels are private messages and public comments. You can see how these relate to other older forms of online communication below:

communications-matrix.png

Social network private messages look a lot like webmail. Public comments on social networks are newer and more interesting. Indeed, Danah Boyd includes public comments as one of the three defining features of social networks (along with Profiles and Friends lists). Unlike message boards, public comments “belong” to a single person and are addressed directly at them. But as Danah has also pointed out (I wish I was half as smart as her!), there is also a performance component to public comments on social networks.

This is best understood with an example. Suppose it’s your birthday, and I know it. If I send you an email wishing you “Happy Birthday” then you’re happy that I remembered. This communication is part of the social lubricant on which relationships are built.

But supposed that I post “Happy Birthday” to your Facebook Wall instead. Then not only do you know that I remembered, but ALL OF YOUR FRIENDS know that I remembered as well. They may find out from the feed, or by visiting your page, but they will know that I’m a good enough friend of yours that I know when is your birthday. That is the performance element of the communication.

Indeed, Danah says that your Friends list is your best guess at the audience for whom you are performing:

The collection of ‘Friends’ is not simply a list of close ties (or what we would normally call ‘friends’). Instead, this feature allows participants to articulate their imagined audience – or who they see as being a part of their world within the site. While SNSes have millions of users, most participants only care about a handful of them. Who they care about is typically represented by the list of Friends. If an individual imagines her profile to be primarily of concern to a handful of close friends, she is quite likely to have very few ‘Friends’ and, if the technology allows it, she’ll keep her profile private. If she wants to be speaking to her broader peers, her Friends list is likely to have hundreds or thousands of Friends who are roughly the same age, have the same style, listen to the same music, and are otherwise quite similar to her. She is also quite likely to keep her profile publically [sic] visible to anyone so that she can find others in her peer group (boyd 2006).

Historically, the Wall (Facebook)/Friend’s Comments (Myspace, Bebo and others) has been the only place on a profile where another user can put something on your page. The rest of the profile has been completely under the author’s control.

However, some of the Facebook apps have changed this paradigm. A number of the most popular apps allow another user to put something on your profile, including #2 Graffiti, #7 X me, #8 Superpoke, #9 Free Gifts, #15 Superwall, #16 Foodfight and lots more. [Note: X me and Superwall are both owned by Rockyou, a Lightspeed company].

In my own experience, performance is an aspect to the use of these apps as well. I feel a certain pressure to choose something “clever” to X someone (e.g. “defenestrate”, “disdain” or “milk”), and if I’m leaving graffiti on a friends page, I try to make it good. The popularity of these apps suggests that social network users are craving more stages for their performances.

I’d be interested to hear what readers think.

Reminder: please switch your RSS feed to feeds.feedburner.com/lightspeedblog – it helps me keep track of RSS readership. Thanks a lot!

Five lessons in viral marketing from a crowd experiment July 15, 2007

Posted by jeremyliew in advertising, Internet, self espression, social media, social networks, viral, viral marketing, web 2.0, widgets.
15 comments

I’ve been traveling a bit this week, speaking at Widgetcon on Wednesday and at Community Next on Saturday. Both panels were on the topic of viral marketing; at Widgetcon with a focus on how brands can use widgets for marketing, and at Community Next with a focus on how to measure viral campaigns.

Dave McClure moderated the panel at Community Next and conducted an interesting experiment with the audience that really encapsulates some of the key lessons of viral marketing. He seeded two memes into the audience. One person was asked to start saying “meep” repeatedly. Another group of five people were asked to put their hand onto another person. The idea was to see which memes spread furthest in the audience.

The “Meeper” juiced up the visibility of his meme by adding an element of clapping as well (“meep”, clap, “meep”, clap etc), and walking up and down the front of the stage. Initially maybe 10 people near the Meeper started to meep as well (and clap – more clapping than meeping actually) but this eventually died down as it failed to get picked up more broadly. The initial early adopters started to feel self conscious when no one followed them, and stopped meeping.

At this point, the people on stage still had no idea what the second meme was until Dave asked how many people were touching someone else. About a third of the audience, maybe 50 people, raised their hand. Although it initially lagged, the second meme had far outpenetrated the first.

Although a somewhat artificial experiment, Dave managed to demonstrate a number of the key lessons about viral marketing in a very clever way:

    1. A “high visibility” app can get quick pickup among early adopters very quickly. “High visibility” can be caused by a high invite rate, inviters who invite a lot of people on average, or simply something that is extremely visible and obvious (e.g. music on your profile page, or some guy walking up and down the stage clapping and saying meep).

    2. High visibility can cut both ways. New users who are seeking social proof can see who is adopting, and decide whether or not they are “like me”.

    3. Early adopters can also be early abandoners and not representative of the broader population. (see Josh’s classic post on the 53,651)

    4. Product matters. While a highly viral app can get distribution quickly, if the uninstall rate is high then it never gets beyond a certain size. While virality dictates the speed of growth, uninstall rate (typically a function of product quality) dictates saturation size, which in many cases is a more important business driver.

    5. It helps to start virality from a larger base. Viral marketing is a probability driven game, and if you don’t have enough initial seeds then a failure of virality from any one seed can stop all growth immediately; with more seeds you have more “shots on goal”.

A lot of the same lessons came out of the panel discussion at Widgetcon. There was a real focus on asking what should be the right metrics for measuring “success” for a widget based marketing campaign. CPM and pixel count (728×90 etc) didn’t seem to be the best way to measure and sell advertising when users voluntarily affiliated themselves with a brand. Echoing lesson #4 above (Product matters), most of the panel came back to engagement with the widget as the key metric for success. If the widget isn’t good, users won’t engage.

As the industry’s attention turns to the tactics of viral growth (whether through email, cross sell, calls to action, optimization of color, font, copy and position etc) it’s a good reminder that, as has always been the case, product matters more than marketing (whether viral or not).

Toyota Prius is the Facebook app of the auto industry July 4, 2007

Posted by jeremyliew in facebook, self espression, social media, social networks, widgets.
7 comments

The NY Times runs a front page article today about the Toyota Prius the hybrid that makes a statement and that sells. The article says:

A riddle: Why has the Toyota Prius enjoyed such success, with sales of more than 400,000 in the United States, when most other hybrid models struggle to find buyers?

One answer may be that buyers of the Prius want everyone to know they are driving a hybrid.

The article quotes a poll that says that 57% of Prius owners cited as their top reason for buying a Prius that “it makes a statement about me”. (Disclaimer: I drive a Prius too).

This behavior is no different than why users of social networks “pimp out” their profile pages with widgets from companies like Rockyou (a Lightspeed company) and Slide. Self expression is a basic human behavior, whether its instantiated in the car you drive, the rubber wrist band you wear, the sticker you put on your bumper, the IM buddy icon you choose or the widgets you put on your social network profile. It sends a message to the rest of the world about what is important to you.

The Toyota Prius is the facebook app of the auto industry.

I would love to hear other examples of online and offline self expression from readers.