The Incredibly Shrinking Software Industry

At one level, Software seems to be entering its Golden Age, quickly spreading into almost every corner of our lives.  TIVOs, iPods, Treos, you name it; they all have a large and growing amount of software in them.  Any yet, despite this explosion of new software, the software industry itself is shrinking.  In 2005, the aggregate market capitalization of the software sector shank by almost 10% despite the broader NASDAQ market being up 1.4%.  Even if one adjusts that number to account for privatizations, M&A and IPOs, the software market still shrank by 9% so there’s no denying that anyway you look at it, 2005 was a bad year for Software stocks.

This shrinkage is also apparent when looks at the raw number of public software companies.  There were 236 public software companies at the start of 2005, but only 213 at the end of the year, a decline of 10%. Put another way, for every new software company that went public in 2005, almost 7 were acquired or went out of business.  Not exactly an encouraging picture.

What then is responsible for the software sector’s precipitous market cap decline?  Like most complex systems, there is no one single factor driving this trend, but a combination of factors including:

1. Software is moving from “growth” to “value”.  Let’s face it: from an investment perspective software is no longer viewed as a “sexy”, high growth industry by many investors.  This change in sentiment has led to what Wall Street calls “investor rotation” as “growth” managers move out and “value” managers move in.  One can sense this change in the daily chatter emanating from the market.  In the past, much of the market chatter was about license growth and new products, now the chatter seems to center around maintenance pricing and services revenues.  More important than chatter though, value managers simply aren’t willing to pay 35X next year’s EPS for anything, which is leading to major multiple contractions in many of the top names in the industry.  I personally believe that this is probably the single biggest technical reason that large cap software stocks performed so badly in 2005.
2. Open Source and SaaS.  The modern software market was built on the backs of large one-time perpetual license sales.  Charging significant up-front fees for access to proprietary source code not only allowed companies to post 95%+ gross margins, but enabled successful software companies to delight investors by very rapidly growing revenues and earnings.  Sure the whole thing crashed when the company ran out of new licensees, but it was a fun ride for everyone in the market while it lasted.  Unfortunately two major trends are conspiring to make it increasingly difficult to grow revenues quickly: Open Source and SaaS.  Open Source basically flips the revenue model: it gives away the source code up-front and tries to make money on the back-end by charging for support.  Open Source has already put tremendous pressure on revenue growth in areas such as web servers, application servers, and databases, and threatens to do the same in several other places.  In response, many traditional “closed source” vendors have reduced their upfront license fees and increased their maintenance charges.  SaaS (Software as a Service) allows companies to purchase software “on demand” over the web.  As a result, SaaS requires little or no up front investment from a customer and is often purchased on a short term subscription plan.  The lack of large up-front payments makes it very difficult to grow SaaS revenues quickly and reduces margins because the company actually provides a real service as opposed to just shipping a disk.  Thus, as Open Source and SaaS gain prominence it’s becoming increasingly clear to investors that the good old days of 200% revenue growth/year at 95%+ gross margins are gone for good and stock multiples are responding by heading south.
3. No big platform transition.  The software industry has traditionally seen its best years in the wake of major “platform” transitions, be it from mainframe to minicomputer, or client-server to n-tier.  Each platform transition is typically accompanied by several new “anchor” products that customers must adopt and this adoption tends to drive new revenues.  For example, client-server led to departmental databases and client-side apps.  N-Tier led to web/application servers.  The industry has very high hopes that Web Services will occasion another major platform transition and thus create a lot of value, but as of yet, no “anchor” products have appeared and revenue growth from Web Services has therefore yet to make up for declining n-tier related growth.
4. Networking companies are encroaching on software company turf. If you pry open the hood of your average router, you won’t find a disk drive or a keyboard but you will find a ton of software sitting inside flash memory or embedded in chips.  In fact, many network company executives will insist to anyone that listens that their company is more of a software company than it is a hardware company.  This really didn’t matter when the “software” that the networking companies built was totally focused on making their devices work better, but increasingly networking companies are focusing on making their devices “application aware”.  To be application aware, these devices basically start to take on capabilities, such as data analysis, workflow, etc., that have traditionally only been handled by “normal” software companies.  Thus, this rise of application aware networking could likely to siphon off revenues from the traditional software industry.  This really isn’t happening yet, but the market is starting to get a sense that this might happen. 
5. Being public ain’t so great.  Being public in these post Sarbanes-Oxley days is not easy, especially if you were a software company that lavished options on its employees and played it a little loose with revenue recognition from time to time.  Like most tech companies, software companies are seeing their margins hurt by higher compliance costs and steep charges for options.  Because software companies have relatively high margins on low revenues, these incremental operating expenses tend to have a bigger impact on overall margins.  What might cost a hardware company with $10BN in revenues 50 basis points of net margin costs a software firm 200 basis points, so the impact can seem to be a bit disproportionate.

Despite these trends, the software industry will still survive and may, at some point, begin to thrive again.  Areas such as web services, mobile computing, semantic analysis, storage management, message aware networking, Chinese software and virtualized computing (to name just a few) hold significant promise for future growth, but that growth will have to occur despite the long-term headwinds posed by the trends outlined above.