Futuristic Play by Andrew Chen

How do you apply math to Web 2.0 websites?

Quantitative approaches to social media
Two great blogs recently on borrowing concepts from the pay-per-click optimization world and applying them to social media sites: How about landing pages for the social media visitor? and also Optimizing social media landing pages.

When I refer to these approaches, I generally refer to a broad set of optimization techniques that advertisers use to make sure the traffic they are paying translates into sales. One example of this is Google's free Website Optimizer tool, which you can use to optimize things like layout, headlines, copy, form placement, etc.

Multivariate regression testing as an example
For example, you could imagine having an invite page for visitors to virally add their friends. Which headline is the best?

• Don't be lonely, add your friends!
• Example your network! Invite friends!
• You have 10 free invitations for friends
• Import your addressbook in 30 seconds

It's hard to tell. And then when you add in 5 different pictures, 5 different explanations of the site, and 5 different color schemes, all of a sudden you have hundreds of different combinations. Well, turns out through multivariate testing you can end up testing just a couple variations and then it'll automatically predict (and test) the best combination of the hundreds.

People in the paid advertising space use these techniques because when you're paying $1 per click (or a $1000 CPM), you try to optimize every opportunity.

Multivariate testing example
Check out this solid example of multivariate testing:

http://dkm.com/

You'll see slight different content, with lots of different combinations of other things.

How to apply hard metrics to social media
Ultimately, it seems like what you want to do is to pick a couple key "transactional" moments that you want to optimize. That could be things like:

• Page where the user creates media
• Page where the user votes or comments
• Page where the user invites their friends
• Page where the user fills out their profile
• Page where the user shares out their widget

That is, a bunch of pages where the users aren't doing the core mechanic of browsing from peer item to peer item. (Like video to video, or person to person). In these moments where people are about to invite friends, you want to figure out the exact right wording to get the person to spread the site virally.

These are the pages where you are most likely to apply these page optimization techniques.

Viral growth dashboard
Also, in a previous blog I had written about 10 different strategies to acquire users. In these cases, you want to build out pipelines and track the metrics around how people are progressing down the goals you want them to accomplish. Here are some possible metrics you'd want to measure:

• Sources of traffic
• Landing page views
• % of users that register
• % of users that send out invites
• # of invites sent out, per user on average
• % of invites delivered successfully
• % of invites read by users
• # of virally added users, per user on average

Or, alternatively, for widgets:

• # of widgets outside
• # of widget impressions
• % CTR of widgets
• Signup conversion rate
• [other viral metrics]

You might also want to track overall growth rate, the viral coefficient, etc.

If you don't watch these numbers, they're unlikely to grow...

Click on this crazy widget, thanks :)

Check out the leaderboard here

How does the online ad gold rush impact your Web 2.0 site?

The stats so far
I love it - the online ad market is going crazy right now. Here's the blow-by-blow so far:

1. First, Google buys Doubleclick for $3.1B
2. Then, Right Media goes for $680MM to Yahoo
3. Even AOL gets into it, buying mobile ad network Third Screen Media
4. And today, WPP buys 24/7 Real Media for $649MM

This means that about $4.4B in transactions in the last month. Wow. That's incredible.

Recall that AOL also owns Advertising.com for $435MM back in 2004, and Yahoo paid $1.6B for Overture back in the day. There's a couple other big acquisitions floating around this that I'm missing.

Uh, what's happening? And how does this fit into Web 2.0?
Unlike the YouTube or Skype deal, where all the bloggers are also the consumers, the ad marketplace is far more complex. It's much less straightforward to understand what's happening other than a bunch of money being exchanged between big companies.

So let's start at the beginning with a history lesson:

Back in the day, you had 3 basic players:

• Internet publishers (Aggregate eyeballs to sell ads against)
• Advertisers and agencies (Companies that want to buy the eyeballs)
• Ad networks (that aggregated websites and advertisers)

Publishers and ad networks were pretty distinct because if you were a big publisher, you'd cut deals yourself. If you were small, then you'd go to one of the many ad networks out there that would sell the ads for you. They stayed separate and advertisers interacted with them differently.

Channel conflict, and the 145 billion dollar mistake
Let's talk about Overture, the company that mainstreamed pay-per-click marketing. Early in their history, they decided that they were going to be an ad network and NOT a publisher. Back then, it was thought that you had to be either one or the other.

The idea is that if they operated Overture.com and got a lot of consumer traction, larger search engines wouldn't want to work with them. So instead, they worked through their publisher channels and relegated Overture.com to a tiny piece of their strategy.

As they grew, eventually their partners realized they were writing very big checks - $500MM or more per year - to Overture, and they grew nervous. Eventually, they threatened to pull the plug on the traffic and build it in house, and one of Overture's customers, Yahoo, bought them.

Again, the concern over the "OR" question - to be an ad network OR a publisher - has to do with channel conflict. If you competed with your clients, it'd be impossible in the long run to maintain those deals. So they chose one over the other.

Google, on the other hand, chose to be an "AND." They both operated a search engine as well as supplying ads to partner search engines, the biggest one being AOL. It turned out that this type of "co-opetition" or "frenemy" relationship was OK in the online ad world! Who would have known?

Turns out this mistake cost Overture $145B in opportunity cost - they were sold for $1.6 and Google is worth about $146.7B today. Whoops.

The jump ball that is Web 2.0
Web 2.0 is a big swath of territory, comprising millions of sites and hundreds of billions of impressions. And of course, all the features that are powering these ad impressions are given away for free. Anybody that supplies the revenue to all of these companies have tremendous leverage - thus leading to the idea of an "advertising platform."

Now that everyone in advertising knows that it's OK to both partner and compete, it's game on.

This means that WPP, a large brand agency that buys from publishers will also operate an ad network that pays publishers. Huh??

This means that RightMedia, a neutral ad exchange that buys and sells ad inventory across many portals, will be operated by one of their customers.

This also means that Google, which pays publishers ad revenue in exchange for real estate on their pages will also operate an enterprise software product that serves ads - which today, charges the publishers money for their services.

So net/net, you can imagine dollars flowing into the online ad ecosystem, circling around a couple times, and then coming back to the company that spent it in the first place. In this ecosystem, the notion of buyer and seller is antiquated. Instead, everyone's just a partner that you might compete with at 9AM, but then will be your best friend at 1PM.

Totally weird, right?

Competition is good for the publisher (That's you!)
If you're operating a Web 2.0 site, this is all good for you. The more players there are, and the more competition you have to buy your ad inventory, the better. Although the ad market is consolidating, it is doing so in a way that many many multi-billion dollar businesses are getting involved - they won't all buy each other, so you won't have to worry about a total Microsoft-like consolidation happening.

In particular, I'm very excited to see WPP play into the market in such an unconventional way. By buying 24/7, they are saying that they won't just be a brand advertising agency, and might instead look at being more of an intermediary. I'm curious how all the other big agencies will respond.

Either way, if all else fails, we can just move to virtual goods :)

Forget advertising - is virtual goods the killer revenue model for Web 2.0?

The breath-taking numbers
First, let's start with the numbers:

• Habbo Hotel generates $77MM in 2006 (Source)
• Club Penguin generates $65MM in 2007 (Source)
• Second Life exchanges $1.MM in the last 24 hours (Source)
• IMVU, Puzzle Pirates, and Stardoll are all doing very well also

These numbers are pretty exciting because they are NOT advertising dollars, but rather people directly purchasing merchandise on website.

Why do ads suck in Web 2.0?
The truth is, most Web 2.0 firms that use Google AdSense or any other ad network typically has a bad experience. When you display contextual ads to people who are browsing for hot girls, you are going against the flow. Ads don't support what people go to social networking sites for.

Ads on people on MySpace are like coupons at a night club.

Because of this, it's well documented that the CPMs for social networking sites are quite bad. Clickthrough rates are very low - in the case of Facebook, you're talking about 0.04%, or 4 clicks in 10,000 impressions. Compare that to Google, which is delivering upwards of 300X the CTR.

Why do virtual goods work in community sites?
Ultimately, virtual goods complement the experience people are looking for in social networking sites. It lets them personalize their online identities more, whether their identities are represented by a profile page or an avatar. Virtual goods can also be exchanged or sold, which complements the idea of gifting or other social expressions.

Virtual goods in social sites is like dressing up or buying someone a drink at a night club

Because virtual goods go with the flow, in terms of what users want and expect, it makes it easier to monetize groups of people. Anyone who gets into an MMO will end up picking up some virtual goods, and by investing themselves, it'll seem natural to spend money.

Where do virtual goods belong in Web 2.0?
Clearly, virtual goods can belong in Web 2.0 in one form or another. After all, MySpace or Facebook could charge points for profile customizations, and you earn those points through actions (like participating in the community) and/or by paying up. Facebook is already trying the virtual goods model, but that's just about icons on the page. I think a better example is Gaia Online, which has done a great job turning a forum website into a full-blown virtual universe.

The really fascinating part about this model is that it support communities, rather than being something that interferes with the user experience. Go with the flow, and money will follow.

My prediction is that these in-browser casual games and online communities will start to converge over time, and you'll start to see Web 2.0 sites start incorporating numerous game mechanics, goal structures, and rich avatar systems.

See you all at Virtual Goods 2007 :)

Happy Mother's Day

Need a solid Web 2.0 business idea? Go upstream from a proven model

Need a solid business idea?
I've recently been asked by a couple people who don't have solid business ideas that they want to work with. In fact, maybe they have something fun and entertaining, but they don't know how to shift it into a model that works.

For all those folks out there, here's the blog for you. Here's a template for how to find a good business concept quickly, using many analogies that are already out there.

Before jumping into this, here are a couple older, related posts that I'd encourage you to read:

• Is Facebook killing dating sites?
• Revenue implications of non-sticky sites
• Eyeballs versus Revenue: What should startups focus on?

Proven transactional models are everywhere
When the Web 1.0 bubble collapsed, some of the only companies left standing were "transactional" sites that weren't dependent on advertising. You could see them everywhere:

• Retail (Amazon, Buy, etc)
• Travel (Expedia, Travelocity, etc)
• Dating (Match, eHarmony, etc)
• Jobs (Dice, Monster, CareerBuilder, etc)

Even these days, you're seeing an entirely new generation of companies focused on tickets, like StubHub, or advertising networks like RightMedia. All of these businesses are real money machines.

How categories have matured, and the battle over users
The only concern for most of these companies, as they scale up, is that their proven business models end up stagnating as other competitors enter the market. Because it's very hard for most of these companies to develop relationships with customers, they end up in a lead generation knife fight.

This means their growth and/or margins whittle down as they engage in ever-competitive user acquisition battles.

Here's some numbers:

The U.S. online dating industry is expected to climb 9 percent year-over-year with revenues of $516 million in 2005 coming from consumer subscriptions alone, said Nate Elliott, an analyst at Jupiter Research. That's slower than the 19 percent growth in 2004. And when compared with a 77 percent jump in 2003, the latest revenue trends seem cause for real concern. (Source)

Online corporate sales will climb to $37 billion. In spite of those numbers, the US online travel market is maturing and growth is slowing down. (Source)

In fact, one of my favorite blogs, Jeremy Liew from Lightspeed Ventures, recently wrote an article on the deterioration of margins within the lead generation industry. (Think online education, insurance, car quotes, etc.)

Where can social media companies fit in?
Given that the real leverage point within these transactional industries is user acquisition, I've come to believe that you can apply social media techniques to build solid businesses that are upstream from the transactions.

The advantages of Web 2.0 sites over the transactional sites are:

• Viral, organic growth in traffic, versus paid user acquisition
• Sticky traffic that can be converted to transactional, versus one-time user visits
• Upstream position from several mature sites means "auction" dynamics

I think the first two are pretty obvious, but let me explain the third one: If you built a site that everyone would visit before they went onto the job sites (for example, a professional social network), then you can "funnel" users to the transactional sites. In fact, you could do one of the following:

• Auction the user off to the highest bidder
• Auction the user off MULTIPLE times, like to the three highest bidders

The latter case is pretty interesting, since if you are able to establish those dynamics, you will have a stable place relative to your downstream partners. They won't be able to pay as much as you can to acquire a user, since obviously Monster would never sell one of their users to CareerBuilder.

Focus on the right social applications
Of course, no one wants to do a social network about jobs, or insurance. Instead, you have to focus on high-interest mass market topics that are upstream to them. For example:

• Retail-> Celebrity gossip and fashion (like PerezHilton)
• Dating -> Hanging out (like MySpace)
• Jobs -> Professional contacts (like LinkedIn)

In each of these cases, you end up with more daily traffic, more unique users, but fewer % of people who are "transacting" at any given time. The idea, of course, is to build a large enough user base that you can a large absolute number of people who are in the transaction stage.

In conclusion, this is a quick way of brainstorming a bunch of interesting social applications that have solid revenue potential. Just look at any transaction that people make over time - be it services, or big ticket items like cars or real estate. Each one of these transactions should have a larger social application to tie them together upstream. By using viral growth techniques, paid advertising, and sticky content, you can engineer a lead generation machine that throws off actual revenue.

Cracking the code: Analyzing viral strategies

Encyclopedia of viral marketing strategies
I've recently been thinking more about viral strategies, and how they get implemented. I've been trying out a couple different sites that have gotten a lot of traction in the last year, to figure out why. After some analysis, it strikes me that there are some things that are obvious to understand, and some things which are not.

Explicit communication is obvious
When you go to a site and they ask you to embed a widget, or import your address book, or something similar, it's obvious to list out these hooks. For the most part, they focus on two sets of actions:

• Assuming control of a communications platform (email, MySpace, IM, etc.)
• Making it easy for the user to spread it (widgets, URLs, Digg this,etc.)

The first case is automatic, yet most of the time they need you to log into an existing service, or download a client software that will do the importing automatically. Either way, it's pretty clear.

Optimization isn't obvious
When you see LinkedIn having an orange button for "Add to your network," it's unclear what other variations that they tried to make the process successful. Was it:

• The placement of the button?
• The color of the button?
• The fact it's a button, not a form?
• The fact it's "Add to your network" rather than "Add a friend?"

It may be that in the universe of thousands of permutations, they are using the most optimal configuration. Yet you'd never know unless you were able to duplicate the same page layout, product, and audience. This makes these types of optimizations very hard to analyze.

Same for landing pages and action pages, where it's critical for some stage of the "user funnel" to be optimized. You'd never know what kind of work they put in to change the title of the page by 10 characters.

Bootstrapping isn't obvious
Another difficult part to understand is how some of these sites bootstrap their viral processes. Obviously, you have to start somewhere. So how widely did they start? Who did they initially court? How were they able to identify likely viral candidates?

For example:

• Buying millions of e-mail addresses and bootstrapping using email marketing (MySpace)
• Sending to all of your friends in your address book (HotOrNot)
• Acquiring advertising media across ad networks (several dating sites)

Obviously in the case of social networks, where you start has a lot to do with where you end. If you start in India, you end up with a lot of international traffic. Figuring out how to break past the first 50k or so users in a targeted demographic is critically important.

Adwords is not enough for success on the consumer web

Viral growth is a requirement for consumer web success
Every web entrepreneur should be a student of organic, viral growth. Even though I've worked in the online advertising industry for the last few years, I'm humbled by the numbers that a well-designed viral site like Tagged can put up. These stats are HUGE, and they are only getting bigger.

Google AdWords is not enough
It's convenient to assume that you can outsource all of your marketing out to Google, and buy lots of AdWords. Or let's say you're ambitious enough to want to buy from ad networks, or other ad sources. This is a myth.

The largest sites ultimately cannot be bought through online ads. To achieve 50 million users via Google Adwords, let's work backwards:

• 50 million users
• 500 million uniques that land on your page (with 10% registering)
• 50 billion ads with a 1% clickthrough rate

Let's pick a range of CPMs, from $0.50 to $2:

• At the low end, $0.50 CPM means $25 million in ad spend
• At the high end, $2 CPM means $100 million

These are obviously impossible numbers to achieve just using ad spend. The only way to solve this is to make sure that every user you bring in brings their 10 friends. Then, when you spend a targeted $10k at the beginning of your site's existence, you can get the momentum going quickly.

Design viral into your product, not into your marketing strategy
A lot of people would claim that once you have a product, you can design a "viral" strategy around it. Do you really believe that designing a word-of-mouth strategy for a music CD is the same thing as a product like Plaxo or LinkedIn? The answer's no. The truth is, you have think about how viral fits into your site at the beginning.

Let's talk about how most people think of viral...

When users see the product, this is what seems to happen:

1. User sees the product
2. User tries the product
3. If user likes product, make it easy to tell their friends
4. Word of mouth spreads
5. New users try the product
6. ... and so on

Let me say that this is a very old-fashioned way of looking at viral product use, because it's inherently tied to an assumption that people have to go through a full cycle of evaluation before they spread the word. What's another way?

How to SHORT-CIRCUIT the viral process

Instead of thinking that viral is something that's tacked onto the end of your user experience, instead think about how it could get integrated into your user process. In fact, the earlier, the better. You want to make it so that in TRYING the product, they spread it to their friends.

Ideally, the funnel looks like:

1. User sees the product
2. User tells all their friends to use the product
3. User tries the product
4. [Site propagates regardless of whether or not the user likes it]

Then regardless of whether or not the user tries the product, or if they like the product, or if they get off their lazy asses to tell their friends, the deed of propagating the site is already complete.

How do you do this? Let's take Tagged as an example - click here to try out the site. Or try Flixster, another site that's gotten incredible traction lately. As soon as they sign in, they are trying to import your Gmail/Hotmail/Y! contacts. They make it an integral part of the user experience.

Or take a look at something like Plaxo, whose entire existence is about asserting control over your communications hub, thus making the viral effect even better.

By spreading the site at the earliest possible moment in the experience, you're not at the whim of the user to act outside of his/her normal processes.

Too annoying? Or too ruthless?
Some might say that they hate the fact that these processes exist, and that they might be too ruthless or even immoral. These points are definitely worth debating. I certainly believe that if you're able to create a engaging user experience, adding these techniques to the front-end isn't bad.

If you believe that:

• Ruby on Rails makes development easier
• Product creation is cheaper
• Infrastructure is cheap

... then the corollary is that there's more competition. Breaking out of the noise in any way possible, even if ruthless, will be a key technique for success in the Web 2.0 world.

If you're interested in a high-level listing of different user acquisition methods, check out an earlier blog I wrote on the subject called 10 obvious ways to ruthlessly acquire users.

Random drawings (that I made!)

Check it out, I actually drew these:

Go to Sketchfu if you want to check out others...

How to fool VCs into thinking you have traction, Part 3

This is the 3rd article in a multi-part blog. Here are quick links to Part 1 and Part 2.

Today, we're going to talk about using automatic page refreshing and other navigational tools to generate extra pageviews for every user, whether or not they are really using the site.

Here's what you say
When people ask you for what kind of traction you have, this is what you say:

"Last month, we got 30 million pageviews - and wow were people engaged. We had an average of 40 pages per session!"

These are great numbers, particularly the engagement factor of many pageviews in the session. These are the kind of stats VCs are looking for when they think about the scarcity of attention. This is particularly true when you show them the hockey stick that's driven by all these sticky pageviews.

If you need an example of why VCs value stickiness, just check out the excellent blog from First Round Capital on the subject, called "Catch and Release Business Models". Basically, their view is that more pageviews per session is a proxy for more passionate users, which means more viral, and better growth, etc.

What's weird with these stats?
Of course, these stats could look great on paper, but in fact they be caused by a number of negative factors which have been documented in the past, particularly two:

1. Inefficiently designed navigation that causes pageviews
   (See the MySpace analysis, called "The Click Factory")
2. Automatic page refreshes which generate garbage pageviews
   (See the analysis on Drudge report in Valleywag)

In both of these cases, users are NOT more engaged, yet generate garbage pageviews that are hard or impossible to monetize. So while a simplistic analysis would just multiple the 30 million pageviews by a comparable CPM, the smart money would figure out how sticky or wasteful the pageviews are, and discount it accordingly.

The second problem of automatic page refreshes is a particular problem for news sites, where it's gray-line justifiable. Oftentimes, this encourages people to stretch it a little further than they need.

On the flip-side, of course, this is the hidden downside of heavily AJAX-y sites that do everything very efficiently. Realize that early on in the process, a site like that might be decreasing their overall pageviews by close to 50%. That'll generate a huge hit to the startup's valuation. It's advantageous for everyone to be educated on this topic, and the pluses and minuses.

How do you figure out the truth?
To understand if these numbers are being inflated, you really have to break down the source of the pageviews on the site. You should ask:

"When users come to your site and expend 40 pages per session, break down what is happening on the 40 pages. What pages are spent on the automatically refreshing news page, versus commenting, versus other activities on the site?"

You typically want to make sure you are separating engagement from the stickiness of the site. Some of this stuff cannot be faked. For example, here are questions that go directly to stickiness:

• How many users come back every day? Every 2 days? Every week? Every month?
• How long do you retain users over time?
• What percentage of your registered userbase is active?
• How many friends do most people have? How many articles do your users forward on?

If you ask these questions, you'll get much more information on how people use the site, rather than relying on a simple metric like pageviews/session. You'll also get a good understanding of how well they can interpret past their metrics into the behavior of the users, and why the users are acting different ways.

Onwards...
In my next article, I'll be covering how people fiddle around with the definitions and standards around pageviews, uniques, and other numbers to inflate their numbers. Stay tuned!