OnStartups

Web 2.0: On Bubbles and Business Models

Posted by Dharmesh Shah on November 30, 2006 in guest strategy startup web20 44 Comments


As I’ve noted on many earlier occasions, a lot of the best content on this blog are reader contributions via their comments.  If I were you, I’d be looking at some of the comments to the more “strategic” articles as there’s usually one or two comments that are really insightful.

This week, in response to my “Commercializing Web 2.0:  Hype vs. Reality” article, I received what has been to date the most comprehensive set of comments from an individual that I’ve ever gotten.  I read the comments over several times.  Thankfully, these were sent to me via email (instead of just posting to the comments area).  I have the author’s permission to share a slightly edited version of the article with the OnStartups.com readership.  

So, our guest author today is Jeff Burchell ( toxic@doobie.com ).  What follows is an only slightly edited version of the email I got from Jeff in response to my article.  I think it’s well worth the read if you are interested in this topic.  My personal thanks to Jeff for taking the time to respond in such detail to my original article.  Thanks also for not leave this as a comment (smile).    [Note:  Jeff and I have never met  nor interacted prior my receipt of his email].

Final note:  The article is somewhat of a rebuttal to my original points (and should be read as such).  Where possible, I changed to points to questions to make things flow better.  But in any case, assume that I “kicked off” the conversation with a point or assertion and Jeff commented on it.  It’ll be easier to consume the article if you keep this in mind.

Web 2.0:  On Bubbles and Business Models

1.  Does More Internet Users = Larger Opportunity?

I don't know that Larger Opportunity is the correct leap of faith here.  While there certainly are more Internet users, they're either foreign (which is hard for a small US developer to monetize), young (don't spend money yet), or not savvy enough to benefit from some of the cutting edge technology.

The most valuable user profile for a new or disruptive technology project is someone who has been online long enough to recognize the value of disruptive technology.  In the earlier days of the commercial internet, the community was a self-selected bunch of people who were interested in technology -- they "got it" much more quickly than does the typical user today.

The exception to this is the teenager contingent, who grew up with technology, but are too young to remember a 28.8k PPP connection (and may not have even heard about the 300 baud BBS days).  These folks are an enormous market, but they're also extremely fickle, and very, very hard to monetize because they don't have much purchasing power yet.

If you are measuring success purely as a factor of page views or unique users, then having a larger population to target may mean larger opportunity.  But the larger online population has driven CPM prices essentially to zero.  If sustainable or profitable business is your measure of success, then the larger, more diverse online population is actually driving your revenue possibilities down, by increasing the amount of users you need to attract before making any real money.

There's a certain Signal-to-noise ratio that's inherent to any sort of crowd.  New internet users are more likely to be in the noise part of that ratio.  So yeah, the pool of people you can draw from is larger, but that doesn't automatically make it one iota better.  Your market might be the same size, and just harder to target among the morass.

2.  Is The Advertising As Revenue Model Now More Viable?

I don't think this is true at all.  There is substantial ad revenue being spent on the internet, but outside of a handful of huge, established properties, it's spread so thin that it's not really a viable model for sustainability.  Part of this is because of the move away from impression-based advertising towards pay-per-click.  Part of this is due to the squatters and ad farms who are gaming the system.  Part is the sheer amount of inventory available that is driving down the payoff per click.  The truth is, to be able to earn more than $1000 a month in pure advertising revenue, you need to be a very popular site, with users that actually act on ads, and/or you need to be able to cut exclusive deals (which means you probably need a commissioned ad-sales person, or someone acting as one -- which adds to your overhead). 

It used to be that if you were pushing 100k pages per day, you could count on enough revenue from Doubleclick and their ilk to pay a small number of people and cover your bandwidth costs.  Today, even though bandwidth costs have gone down, you need to be pushing ten times that number of pages to get anything other than a meager check from AdSense or YPN.  Bandwidth costs have not gone down by a factor of ten.  Technical or other hands-on staffing costs haven't gone down either.  If you're making $100 a month in advertising revenue, you have a hobby, not a business.

AdSense has brought many more sites the ability to collect tiny checks, but it's done so at the expense of the midsized sites (you know -- the ones that had proved their models, were making some money, and were in the best position to go public if that was their strategy).  To get to the point where you're serving 30 million unique users a month, you've had to go through many months of serving 1, 3, 5, 7 million -- where your costs are increasing, but your revenue possibilities are still crap.  The deck is totally stacked against organic growth, and if you can't make the leap from 1 million to 10 million exponentially, you're in big trouble.  The way to grow exponentially is through buzz and hype -- but organic growth is more about listening to what your users want, and actually giving them the valuable experience they're looking for. 

Which one of those scenarios is more "valuable" depends on how you and the marketplace define value.

The fact that Google and Yahoo own the ad networks has got to be a part of this.  By limiting the type of growth of midsize sites (either intentionally or coincidentally), their position at the top of the pyramid is safer.  And when someone threatens to get too big and actually become legitimate competition, well, then it's time to call in the M&A; team.  I'm sure that it doesn't hurt one bit that the ad-networks are probably in THE BEST position to really know what a site's traffic is, when compared to other sites (since we all know that Alexa sucks).



3.      Can Economics Of User Generated Content Work? 

Creating content that drives a sufficient a significantly sized audience is still expensive.   GOOD user generated content is still not produced for free.  An obvious example of this is with the Podcast space and with "user generated" video sites like YouTube.  For every viewer of a vlogger on YouTube, there are tens of thousands of viewers of CBS's repurposed Late Show clips, or bootlegged clips of The Daily Show or Keith Olbermann's most recent rant.  These are all professionally produced shows, that cost thousands of dollars to produce (at union rates, no less).  Even lonelygirl15 is "professional" (in that she's a paid actress, and it is her and several other people's dayjobs).

CBS recently announced that its use of YouTube has raised ratings among its traditional markets.  The same is definitely true of MSNBC, though to my knowledge, they haven't published those numbers (but I'd bet that few people knew who KO was before seeing him tear into the Bush administration online, and some of those people have certainly become viewers).

CBS has figured out something obvious.  They could be hosting these clips on their own servers, on their own dime, and spend marketing dollars to tell viewers and non-viewers that the clips are there.  But youtube not only lets them host them for FREE, but brings them at least some audience for free.  That's marketing money that CBS doesn't have to spend, and now that they've seen the positive results of it, you can expect to see the other networks following suit. 

The same thing is happening at MySpace.  It once may have been a place for friends, and most of the profiles were real people, sharing real things.  Now, you can be Burger King's friend.  Over 100,000 people are.  Are there any profiles that are truly User Generated that have 100,000 "friends"?  The same thing is going to happen to Facebook any day now (users there keep getting invited to join "groups" that seem marketing-driven, if not downright spam).  Blogger is already a lost cause, with all the splogs.  and so on and so forth.

True UGC sites are small, serve niche markets, and don't make much in the way of revenue (think BBSes).  As soon as the large dollars start to flow in, the focus of the site shifts, and the non-paying users become less important than the paying users -- which completely breaks the UGC model away from "users are king" back to "content is king". 

And that's where it falls apart.  The truth is, good content is not easy to produce, and professionals will consistently outperform amateurs in writing or otherwise producing content that people actually want to consume.  There will always be exceptions to this, but the highly skilled amateurs tend to evolve into professionals (witness what happened with the blogs... we now have a handful of A-List bloggers who make their living as writers, and are receiving most of the revenue that's available to the blogosphere)

The merger of this is what makes reddit, and digg, and sites like metafilter so great.  You've got users (on some sites, smart users) that can help to steer you towards professionally created content that they find to be interesting.  You can use the wisdom of the crowd to find the best of what's out there.  That's different than amateurs producing better stuff than pros.  Call it user rated content, not user generated.

4.  Is There A Big Technology Disruption In Web 2.0?  
 
Oh, there's still disruptive technology.  It's just not really on the Web.

Sonos is disruptive.  Sonos plus Rhapsody is hugely so.  It's the fabled "Jukebox in the sky" realized, with a business model based on established rules of selling hardware at a small profit, and collecting ongoing subscription fees for a service that isn't available through other means.  Just wait until they start adding pandora-like technology to it.  Pray that they don't add payola-type technology to it.

VoIP wants to be (and seems to be getting a little more momentum, especially as we start to see hybrid devices, like cellular phones with WiFi, and landline-replacement devices that speak both POTS and VoIP.)   While Skype and futurephone and jajah are driving long-distance costs to zero for people willing to jump through some hoops, there's progress being made in making those hoops easier to navigate.

Video across IP is going to be a threat to cable, _IF_ the content producers are willing to play (which might mean it gains traction overseas before in the
US).  Hundreds of thousands of people are already downloading TV shows via BitTorrent (jumping through hoops).  If someone made it trivially easy (and legal), they'd have a winner on their hands.  And if the TV networks saw it as another distribution channel (especially one for their mostly unused back catalogs), the viewers would love it.  But the key is content.  If Video over IP only has shows worthy of cable access, it's doomed (or it'll remain the realm of pirates), but if it's used as a means to (cheaply) distribute content that's already being consumed via more traditional media, it will only expand the audience, and give the users what they want -- more choices, more convenience, more control.


The NFL offers the Sunday Ticket package to US Satellite viewers.  For a couple hundred bucks, you can have access to all NFL games on any given sunday.  If you're not in the
US, you can spend a couple hundred bucks and stream the games from the net, at virtually the same resolution that you can get from the satellite.  This is not heavily marketed, and it's clearly a pilot program, but it appears to be successful.  Apple is claiming some success selling sitcoms on iTunes (though why people watch TV on an iPod is beyond me).  If they can get their TV hardware out the door, they're going to increase their sales exponentially -- now if only they could get away from the pay-per-use model and into a subscription model.

At the same time, there's hardware on the market today that does similar things (Netgear's eva700 comes to mind).  That device, with a fast enough connection, and some upstream support, could easily replace Cable TV or DirecTV and your video rental store.  Build it like the Sonos/Rhapsody system, charge a subscription fee, and you've got the ultimate Tivo (so wait, when I sign up for a season pass, I can get all the past episodes too?  Cool!), and an big fat threat to the cable monopolies (which the consumers hate anyway). 

Apple doesn't usually announce vaporware, so the preannouncement of the iTV is out of character for them.  I think they're scared of losing this market, and trying to convince the content providers that they will be in the market soon trying to keep them from signing deals with other vendors.  That's likely to make it difficult for a startup to compete fairly.



Point is, the “web as platform” is a powerful concept, but the implementation is still very, very early and nobody has really figured out how to divvy up the revenue that is generated (if there is any revenue).

Web as platform is a non-starter.  If I write a mashup using data from three sources, I'm relying on the reliability of all three of those sources (which opens me up to a lot of risk that I have very little control over).  Given that I may be competing with one or more of them (if I make their data more valuable, I'm probably taking some of their users), they don't have incentive to continue providing me with data reliably, unless they start charging me for the data.  But, we've come to expect that data on the web is free or ad supported.

Witness the Cease and Desist that was sent from Google to the Gaia project folks over the use of the Google Earth data.

5.  What Is The Impact Of Better Marketing and Distribution? 

This is the single *best* thing about "Web 2.0".  Smart people have learned that marketing isn't black magic, won't give you "hockey stick" returns by itself, and doesn't require someone who wears well-marketed $1500 shoes (and demands the salary to keep buying them).   If we'd known this in 1996, the first bubble popping might not have been so painful.

Good marketing is the creation of hype.  Good business is being able to realize the expectations of the people who hear the hype (and making sure your marketers are skillfully managing the hype that people hear, rather than just trying to generate it louder than anyone else). The first bubble was fueled almost purely by unchecked hype -- and a lot of that was driven by marketeers, who were often rewarded for it. 

A year ago, the Web 2.0 hype factor was much lower than it is now (and it appears to be growing).  If we don't get it back into check, we're totally doomed.



6.  Capital Efficiency:  It is now possible to get a new startup off the ground without raising any institutional capital.  

Which means it's possible to get a startup off the ground without having to defend your ideas and intentions in front of people who have heard lots of other ideas, and know how to choose the good ones, and can ask the questions that help you better refine your good idea into an awesome one. 

This one's a two-way street.  Sometimes the point of raising money isn't what the money buys, it's what the money brings with it, in the form of the rolodexes of your VCs, or the recommendation of a good infrastructure guy, or the validation of being selected out of a large group by a firm that's known to be selective.

Look at the often mentioned YCombinator.  Most of the people selected for YC are students at prestigious, expensive universities.  They're sending in ideas to YC to get living expenses for a few months -- but most of them already have access to that kind of capital (dude, ask your parents, or your friends' parents, or your rich uncle, or your old boss, or sell your car and buy a beater).

The thing is, it's not the money that YC gives you that's valuable.  Sure, it'll rent you an apartment and buy you a laptop.  There's some value in the business services that they provide (lawyers aren't cheap).  But what's really valuable is the name recognition that you get by being a YC-funded startup, and the people that Paul Graham and rtm can introduce you to.  There's no way we'd have heard of loopt already if they'd been bootstrapped (especially since we're so far outside of their target market).  There's no possible way that Reddit would've been sold before digg -- they'd still be building name recognition.


7.  Tighter IPO Market:  Unlike the frenzy we saw during the last bubble, there is no longer a pool of retain investors sitting on the sidelines waiting to jump on the next hot technology stock. 

Right, there's a pool of larger companies with currency in the form of their stock.  M&A; is a very viable exit strategy (and always has been), and now that costs are lower to get a company off the ground, it's still possible to get a 10x or 50x return in a few years.  Where it used to cost $50M to build a company that could go public with a $500M market cap, now you can build a company for $10M and sell it to one of the big guys for $200M (or more).

That's still a frenzy.  It's just one with a different exit strategy.  And it's showing many of the same flaws (especially now that the returns are getting bigger).  Back in the dark ages, there were countless business plans that pointed at previous successful IPOs as validation that building a company with a negative revenue stream but interesting potential was a good bet.  In some cases, it was a good bet, but in most cases it wasn't.  However -- the validation of someone else's success, and the crazy groupthink going on at the time made it look like a much safer bet than it was.  And we're seeing that repeat itself.

Before the YouTube acquisition, the articles about YT were all about how unsustainable their business was, because they were spending a fortune in bandwidth serving video that was mostly violating someone's copyright.  And then, poof, Google comes in and saves the day (at a higher price than YT could've gotten had they tried to go public).  The acquisition didn't change YT's business model, but in many people's eyes, it validated it.  Suddenly, all the ink is about how ingenious YT is, that it shows we're in the middle of the next revolution.  It reads a lot like the ink before and after the Netscape IPO (or about Napster), and we know what happened there.

Now, people are using that success to justify spending a fortune on bandwidth, expecting someone to come in and save the day through a buyout.  While that's a great exit strategy for the few companies that can pull it off (which is as much a factor of timing, connections, and luck as it is pure merit), that's not sustainable either.

And while it helps the founders and investors, unsustainable business practices leading to a high-dollar acquisition doesn't actually help the business itself.  Just ask @home, Netscape, Hotmail, Inktomi, Go, Geocities, even Hotwired (which became Lycos, and has since all but faded).



So, what do you think?  Putting aside whether you like the term “Web 2.0” or not, do you think we are in (or entering) a new Bubble?  Or, have enough of the dynamics of the Internet industry changed to sustain some of the patterns we are now seeing these startups exhibit? 

We are well into a new bubble, and it will either start deflating in a controlled manner very soon, or it's going to pop in a spectacular fashion.

One of the biggest ways that I know this to be true is because of the veterans of the first bubble, and what they have to say from inside the bodies of the companies on the edges of this one.  I know people inside most of the big name 2.0 companies, and essentially all of the top 10 big-name 1.0 sites.   Most of the folks that I know, I know from the Web 0.95 days (and we were the ones that were still working during the lean times post Nasdaq-crash).  We've seen it all, many of us knew the first crash was coming, and most of the people that I hear from are seeing the same things happening again.  (Many people are reporting that lots of the people responsible for making the same broken decisions are NOT people who were inside the industry the first time around.)  The mistakes aren't as visible from the outside, perhaps (not as obvious as million dollar launch parties or 500 aeron chairs), but they're based on the same broken logic.   Product, ?????, Profit is no better today than it ever was.

People starting businesses today aren't talking about real revenue streams, or how they're going to build value (or even how they're going to solve a real problem that people really have).  They're talking about how they're going to build a company that'll be a target of an acquisition in 18 months.  That's bubble thinking.  In bubble 1.0, it was how to go public in 18 months.  Now we've learned that those days are over, but these days, there's money to be made selling your company to a public company, rather than directly to the markets.   (And remember, this is coming from someone who made some money selling a startup to a public company)

And thus, the bubble becomes self-inflating.  Those who got in early generally were solving a real problem in a new way and are now getting out with fatter and fatter checks -- and rather than being celebrated as a combination of innovation, luck and timing,  this is being used as a justification for more people to get in (and for more companies to get funded).  This draws in some that're capable, but risk-averse (like a VC firm that's ready to spend $2M on an idea that they wouldn't touch at $10M, or a sysadmin with a mortgage and baby who'll take a $80k salary, but not a $10k one plus equity).  Again, there's some success and some failures, but it's the success that we mostly hear about.  And then, the fact that so many people are all on the same (self-fulfilling "wisdom of the crowd") bandwagon becomes the justification that the bandwagon is not just a fluke, but a new way of doing business -- so people start getting in without really understanding the true dynamics.  And then, pop!

We're probably now entering the second half of "people think this is a new way of doing business" phase.  When the MBA students are building models based on monetizing a wiki (and they are),  and then actually expect to join a firm creating commercial wikis immediately out of school (or better yet, starting one -- despite having no experience with writing software or bringing it to market), you know the days are numbered.

I'm an unrepentant capitalist, and that's treated me well over the years.  If you can build a business that, once established, slowly earns money, and earns more with each new customer, then your limiting factor is your customers -- if you can build something that has sufficient appeal, you will eventually retire rich.  But if your business is one that isn't directly monetizing your customers, such that each new user actually loses you money (like YouTube and their bandwidth bills), then your limiting factor is your burn rate and the depth of your pockets.  If you can balance your appeal to users against your appeal to a potential acquire-er, perhaps there will be an exit event that will make you instantly rich (though the odds are very much against you).  But with each exit event, more and more people delude themselves into thinking that "It could've been me" and get fantasies of absurdly large payoffs.  It's that thinking that inflates the bubble (and increases the noise, making it less likely that your awesome product will get the attention it needs), and I think we're seeing more of that thinking than true innovation right now.

And in the end, it's innovation that wins.  Always.



Thanks again to Jeff for taking the time to share his ideas with the OnStartups readership.  If you have counter-points to Jeff’s, please leave comments and keep the discussion going.