During my career I have had the privilege and the misfortune to be involved with quite a few nascent businesses. Some have been small, with no more than five or ten employees. To this day they still have that many or less. Others grew from 50 employees to over 10,000 in a very short time. In every case there was a technology component involved in my job.

During this time I have consistently been an IT buyer. I still cringe at the $3,000 a piece I spent for DX2-33 PCs with 8MB for a financial services firm back in the mid-1990s, though at the time it was a heck of a deal. I still fondly remember the first time I successfully installed Linux from floppies downloaded for free from the Internet onto those PCs, though both of them were rewarding acquisitions for different reasons.

Not so many years back I was involved in a start-up. We had tons of money in the bank from a generous VC and a bright future ahead of us. After all it was 2000, the height of the dot-com bubble. We were instructed by our board of directors to build out infrastructure, hire employees, and be ready to "scale." Being a simple country boy, this didn't make a lot of sense to me, since I was raised on the aphorism, "If you don't have it, don't spend it," and other lessons of frugality. At the time I figured that those older and wiser than I were giving sage advice. Let's just say that chapter of my career didn't end as lucratively as others. I suspect there are many others who had similar experiences during that time.

Fast forward to present day - I am involved in another start-up. We sit four to an office (some days we cram in more). We watch our pennies and cash flow and tighten our belts when things start to drift out of line. In the heyday of the tech bubble, companies like ours bought full-page advertisements in thick glossy trade magazines (many of which have since folded or moved to online-only publications). Today we shrewdly buy online advertisements that we can update in real time and we post to blogs to promote ourselves.

The most notable change between then and now is that I don't invest capital in infrastructure to the degree I once did. In 2001, I had a beautiful fully redundant data center with shiny new racks, raised panel floors, and the snazziest ultra-secure firewall. Today, we lease servers and use hosted infrastructure at reasonable monthly fees and we can choose to migrate to another provider with very short notice. Also we don't tend to buy software per se; we buy service contracts or hosted solutions. In my early days I purchased, as a percentage, a large number of Microsoft products; today I use far fewer.

I don't begrudge Microsoft its fortunes. Really I don't. I also doubt that we will ever see an "open source billionaire" and I wonder if we will see a new proprietary software billionaire in the near future. I think that time has past. The model of build it once, sell it 100 million times (or more) is probably not going to work for too many software vendors. Times are changing; as consumers of technology we require more flexibility and adaptability that match the pace at which we do business. Going forward, the most practical way for a software company coming to market these days to achieve widespread success is to not only make a good product but also leverage the efforts of others. If Microsoft and other traditional proprietary vendors lower the barrier to use and improve their ability to integrate their software with others, I probably will buy more of their products.

Let's use a non-open source example to prove my point - SalesForce.com. When they launched, Customer Relationship Management (CRM) software had become expensive, it required a full-time administrator to customize and keep running, and feature requests were greater than any one software company could adequately address. Along came SalesForce.com with a hosted model that virtually any small business could adopt for a relatively small price. Plus they opened their AppExchange where others can offer additional add-ons to their product, making their product very sticky by adding features in a leveraged manner. Ironically SalesForce.com is no longer being used in just small businesses; they now have customers with thousands of users.

The change in the software industry is not simply the rise of open source software; it's the need of the customer to maximize the value that she gets from the new technology. In response, software offerings are changing, and vendors are obligated to continue to earn the loyalty of their customers. This model is one where software vendors are beginning to resemble service providers. They are obligated to improve and service their customer or risk losing them. As I look at the old days versus today, there's a big difference.

The Old Days: On-site data centers, servers, T1s, firewalls, and OS software
Today: Leased servers in someone else's data centers

The Old Days: Commercial firewalls hosted on commercial Unix and proprietary routers
Today: Inexpensive SoHo routers flashed with embedded Linux and complex security and networking tools

The Old Days: Commercial PBX phone switch, with voicemail system for tens of thousands
Today: Asterisk-based phone switch for less than $1,000 a year

The Old Days: Solaris or Windows NT Web Server running Netscape Server or IIS
Today: Debian or Red Hat Linux running Apache

The Old Days: Lead tracking in commercial CRM solutions with proprietary databases
Today: Open source CRM or hosted CRM solutions using open source SQL databases

The Old Days: Solaris, Windows, and Oracle
Today: Linux, MySQL, and PostgresSQL

The Old Days: Large capital outlays, yearly and maintenance contracts
Today: Low upfront investment, subscription fees

Despite these examples I still buy things: Dell desktops running Windows XP or Mac Book Pro Microsoft Office. I am sure our accounting guy runs QuickBooks or Peachtree or some other commercial accounting package, though I do believe we use a hosted version of that software with online access. When given the choice though, I tend to go for the option that has a subscription, which reduces my upfront expense and doesn't lock me in.

I think there is a phenomenon in enterprise software that goes something like this. You buy an application for a large chunk of change and once you get it into production, it just doesn't meet your needs. In order to get the appropriate ROI (or save face in front of your CIO), you keep that system for a number of years despite the fact it never works as promised. In the meantime there is an opportunity cost to your implementation and little empirical data to support the losses in productivity.

It's easy to dismiss the open source community as a bunch of overly enthusiastic technophiles. I see them as people who intelligently identified and embraced and helped to create this new trend. The open source community is a group of individuals and companies that are effectively harnessing the power of peer production to develop high quality, feature-rich applications that meet technical needs. This development model also lends itself to greater flexibility in product features, distribution, and releases. While the attention getter for open source usually is that it's free, it's not the real hook - flexibility is. The trend is that buyers want flexibility, including price and licensing terms. That's why open source and Software-as-a-Service are increasingly prominent in my and probably your IT purchases.