The three rules of real estate are widely known: location, location, location.
The four rules of venture capital remain a mystery to most: too early, too early, too early, too late.
Case in point: digital media. A number of funds were formed to do digital media deals a decade ago during the Info Superhighway craze that led to Big Telcos romancing Hollywood. As they say, a shotgun wedding. It ended badly, and these funds performed poorly. One wants to invest when the trend is clear, but before the herd drives value up.
Are we there yet with digital media? The answer seems to be yes. It will be seen as the second Internet Bubble, and should be fast and furious once underway, perhaps as soon as 2007-8. To understand why, consider the Double Bubble phenomenon of technology innovations.
Tech developments run on a different clock than stock market trends. In 1978, we were in decaying economic conditions. Stagflation. High interest rates. Second oil crisis brewing. Jimmy Carter's malaise speech. Stock market heading south, especially in inflation-adjusted (real) terms. Yet 1978 marked the beginning of the PC Bubble, a glorious five years that set the true foundation of venture capital as a business rather than a cottage industry. The microprocessor had come of age, and investments in PC companies quickened. Valuations rose. Companies with names like Kentucky Fried Computers got funded. Apple went public in 1980 at the then unheard of market cap of $2B ($5B in 1995 Dollars) just as the economic crisis hit its peak. Gold at $800. Inflation over 10%. Interest rates to double digits. Stock market to a low in real terms close to the lows in the 1930s. Double dip recession. Yet the PC Bubble continued. The IPOs in 1983 were fast and furious. Then the Bubble burst. The last to get out was the aptly named Fortune Systems. The founders made a fortune, the investors got fleeced.
The rhythm of tech has been driven by the pace of semiconductor improvement. Moore's Law gives a 100x improvement every decade; systems and software provide a 10x improvement to the end-user. Captured in the well-known aphorism: "Intel giveth, and Microsoft taketh away." Transistor is invented in 1947. Integrated circuit in 1957. Large Scale integration in 1968. Microprocessor (8 bit) in 1976, with the seminal Intel 8080, Zilog Z80 and Motorola 6800 all coming to market in 1978. These improvements all led to tech bubblets. The Sputnik defense pop in the early '60s. The minicomputer boomlet in the early '70s (Kleiner Perkin's first fund had one hit, Tandem, from that era). And the PC Bubble in the early '80s.
A decade later, in 1986, Intel perfected the 32-but microprocessor. It did not spark another boomlet, although it did lead to the rapid death of DEC, Data General, and IBM (as a technology leader). Instead of spawning a new tech boom of start-ups, value flowed to the leaders of the PC industry, particularly after 1991 when Microsoft launched Windows 3.0, leading to another aphorism: "Microsoft doesn't get it right until the third release." Masayoshi Son, the founder of Softbank, says he made his first fortune during this period by investing in the top 40 PC companies that came out of the bubble. Around 20 made it big, and overall delivered a 40x return, better than anyone made in the PC Bubble itself.
In 1995 came the Pentium. More importantly, with it came faster video and communications processing. The decade clock of tech had produced modems fast enough to enable the Internet, and graphics fast enough to make it visual. The Internet had been around, but had been based on a dreadful command-line interface and obscure search methods called Gophers, Veronicas and Archies. Then along came a brilliant synthesis of the pieces: the World Wide Web, with HTML as the foundation for a graphical interface. We had the first Internet Bubble. Netscape went public at the then unheard of value of $5B, mirroring Apple 15 years before. The dot-com era was on, and new franchises emerged. Companies with names like Pets.com got funded. Ann Winblad, a venture capitalist with vague ties from the PC Bubble to Bill Gates, was quoted as saying she knows the valuations are nuts but she feels compelled to invest anyway. (It never seems to amaze that even if we see the bubble we stampede with the herd anyway - real estate anyone?) As the bubble popped in 2000, we had around 200 Internet leaders, 5x more than came out of the PC Era, with the expectation that around 100 would thrive.
Now the decade clock has ticked again. The faster chips are being applied to embody incredibly complex algorithms for video compression and wireless bandwidth. We are approaching a point where the 100x improvements in chips can lead to a 10x improvement to the end-user: broadband, WiFi, 3G, and HD. The pieces are starting to fall in place. The bandwidth algorithms allow faster broadband over both copper and air, in the form of VDSL, WiFi and soon WiMAX. The video compression algorithms allow download of high-definition movies over the now speedier broadband pipes. New digital rights management technologies allow safe distribution of high-value digital content over the wild and wholly worldwide web. The dramatic success of the combo of iTunes and iPods has shown the potential for digital download of protected content when the process is made easy enough for normal consumers.
Are we there yet? Will the Broadband Age spawn yet another bubble that creates new franchises out of start-ups, as happened in 1978 and 1995, or, will value flow to the Internet leaders, as happened after 1986 in the PC Era? The answer is mixed but promising.
One thing is clear - value in broadband will not go to the PC leaders. They are yesterday's news. The Wintel hegemony is mature, if not dead. We can see this in the S curve for R&D: how much investment must be made (x axis) for improvements in price/performance (y axis)? The S Curve that results reflects the process of innovation. After an unpredictably long period of core research, a product emerges, and modest investment leads to rapid improvement. That is the knee of the curve. Eventually the curve flattens, and massive investment results in modest improvement. This is not a problem per se for the leaders, as the barrier to entry is insurmountable - until something new emerges. We are at the point of a flattening S Curve with Wintel. Intel has backed off continuous improvements in the Pentium chip, as diseconomies have arisen - too much heat, too much complexity. Instead Intel has shifted to multicore processors on the chip, and the best chip minds are well beyond the processor into complex algorithms for RF and compression. Microsoft is mired in filling security holes in Windows, and has delayed and dumbed-down its Longhorn release. In both cases, $Billions are invested for minimal improvement. To top it off, something new has emerged - the cellphone as the next PC. Like the mainframe, the PC lives on, but the game has shifted.
Broadband value has already begun to go to the Internet leaders. The first market bottom in 2003 after the crash saw a revival of high P/E's in the Internet leaders such as Yahoo, Amazon and eBay. A plethora of Internet ad ventures emerged out of the rubble of 2000. A new leader emerged in Google, and much of the value increase of the Internet since 2003 has now shifted to that single company, much as Microsoft swallowed the lion's share of value after 1991 in the consolidation of value from the PC Era. Yet no matter how voracious is Google, it is in a relatively narrow slice of the digital media pie.
A glimmer of the next bubble lies in Jamdat, a mobile games company which has gone public and sustains a 10x multiple to revenue, as bubblicious as an investor could want. The appetite for the new swath of digital media ventures is there, but the number of candidates remains quite slim. Are we there yet?
The Internet Bubble was of a different character than the PC Bubble. Not only was it larger, it paid for the buildout of a network technology, and joins the classic bubbles of American Innovation over the past 200 years. The first American bubble was the canal bubble of the 1820s. It led to wild speculative swings and the Panic of 1837, resulting in the first Great Depression of the young republic. The second bubble was the Railroad Bubble of the 1870s, which led to the Panic of 1873, was followed by a second RR bubble (a Double Bubble back then) that led to the Crash of 1893, causing the second Great Depression in our history. The third was the automobile (highway) bubble of the 1910s, which started with the Henry Ford bubble of and a subsequent 72% drop in the Dow, and then of course the GM bubble of the Roaring 1920s, the Crash of 1929, and the third Great Depression. (As with the self-acclaimed Greatest Generation, this was the Greatest Depression.) There is something about network buildouts that they come with a speculative bubble, creating overcapacity fueled by debt, leading to a stock market crash, and subsequent workout of the debt. Perhaps it takes a mania to overcome the normal inertia of government and big business to change, and cause an otherwise sensible company like Time-Warner to sell out to an upstart like AOL at values only a Mother of All Stock Asset Bubbles could love. 
It is well known that innovation comes into a market via an S Curve. What is less well known is that investors react to the S Curve with two bubbles: the Hype Bubble when the technology trend is first recognized by the herd; and the Growth Bubble when the growth is then recognized as continuing. The first bubble causes investors to get way ahead of the market, and when they realize this, it corrects; the second bubble comes when they begin to realize that indeed the early promise is being realized, and they jump back in. The tipping point appears to be when the penetration of the network technology reaches 50% of its ultimate market.
These historical network bubbles all show the recurring pattern of a Double Bubble. First you build the RR, then you ship the goods. The first RR bubble led to dramatic price drops in rail service; this spawned a gaggle of golden gooses to emerge: Pullman cars, express delivery (the FedEx of the day), refrigerated cars (to move beef from the Midwest), and the greatest gander of them all, Sears, which began as a mail order catalog to deliver virtually anything by rail to the great Midwest.
A similar pattern is beginning to emerge from the Internet Bubble. Cost of broadband has plummeted, and new services have begun to emerge, both over wires and over the air. The Internet hit the knee of its curve in 1995 and has grown remarkably quickly ever since. The stock market, however, got ahead of itself, and we had the first bubble from 1996-2000. When the market corrects, it overcorrects, and interest wanes in the category. As happened here. Yet the underlying business trend continues. At the point of 50% penetration, the herd once again recognizes that the trend has continued, and jumps back in. This is the second bubble.
We hit this point with the Internet portals in 2003. After the crash, there was considerable second-guessing of the bold predictions of growth that even Time-Warner swallowed. Analysis in 2002 indicated that indeed those predictions had been fairly accurate, but with a one year lag. The trend was continuing. This caused a return to the former leaders and the 2003 pop in value that led to the Google IPO. Eventually even the popular press recognized this phenomenon. As The New York Times put it, It's
To venture investors, this means there are two sound times to invest:
- early, after the trend but before the herd; and
- late, after the Hype Bubble pops but before the continuation of the trend is recognized.
Joining the herd is a (greater) fool's errand. What bedevils venture investors is often it takes a longer time than expected for the new technology to hit the knee of the curve, at around 10% penetration, and sometimes the technology fails to take at all (pen-based computing, anyone?). On the other hand, once it takes off, it is a tornado that is hard to resist.
Inside of the large Internet S Curve, there are smaller topical curves: broadband penetration; wireless penetration; and (coming) digital TV penetration. Wireless hit 50% around 2004, and the Jamdat IPO was well-timed. Broadband penetration is soon to hit 50%, and we should shortly expect a flurry of attention on digital download of media, beyond the iPod. Soon after digital TV will hit its 50%, and the next wave will be off and running. While in the broader scheme, these curves will all blur into a general Broadband Penetration curve, the timing of the specific curves inform us as to where to focus attention and when.