My first job was with IBM in sales, and that gave me a great grounding in business. But I really became enamored of starting my own company when I read a fascinating article in 1976 about the invention of the microprocessor, and the potential for what it might mean for the then largely mainframe dominated computer industry.

I got my main chance to join this exciting new world when in 1981, I ran into a little company then called Relational Software, Inc., which had developed a database software package called Oracle for early DEC minicomputers. The entrepreneur running that company offered me a job as one of his early sales and marketing employees.

Well, one thing led to another, and I was soon hired as employee #17 at Oracle, working for Larry Ellison, and writing Oracle's first formal business plan. Shortly thereafter, I became a vice president in charge of Oracle's first forays into the microcomputer and PC business. I learned a lot about starting and building a business from Larry, and I will forever be grateful for that. My entrepreneurial career over the next 30 years would not have been possible without the three years I spent at Oracle learning my craft from one of the most successful entrepreneurs of our time.

Three years after I joined Oracle, I spotted an opportunity to start my own company, and I took it. The PC had been introduced in 1980, but notwithstanding its designation as a "personal computer" it was largely being used as an office productivity tool to replace typewriters, calculators, and filing cabinets. What made the early PC so successful was software companies of that era who made word-processing, spreadsheet and database packages. Local area networks (LANs) had just been invented, but very little software existed to truly take advantage of these PCs all hooked together in a corporate setting.

So in 1984, I left Oracle to co-found a company called Gupta Technologies that built a SQL database server for PC LANs. Our Gupta SQL System software also included an application development tool for Microsoft Windows and SQL connectivity software. We called that type of network configuration "client-server software," and by the late eighties I found myself helping to usher in what came to be known as the client-server computing revolution.

Building and running Gupta Technologies was an absolute blast for me. For the first eight years after our founding, we doubled each year with very little venture capital funding. We were soon considered one of the hottest companies in the software industry and took our company public in January 1993. For a while we thought we would be the next Oracle. But that was not meant to be. Very quickly, the entire client-server tools business became a crowded market with many new entrants, Oracle modified its successful mainframe and minicomputer database software to run on PCs, and Microsoft got into the business with its own Windows based SQL Server product.

And then the Internet came along, which represented an even bigger blow for our business. By 1996, the ideal corporate application had changed from being "client server" to "Internet based." The technology revolution I had helped to start was over, and I didn't have the heart to carry on any more. I decided to sell all my shares in Gupta Technologies and to leave the company. Before leaving, I even changed our corporate name so I could separate myself from the company I had founded.

It was the hardest thing I ever did. For the first time in my life I felt like a failure and I didn't know what I was going to do next.

During the next year, I went through a deeply introspective period and I became I obsessed with figuring out what I'd done wrong. I vowed to myself that the next time, if there was going to be a next time, I would build a "business to last." My experience at Gupta Technologies taught me the first and most important valuable lesson of my Silicon Valley career: that technology is a ticket to the game but not the game itself.

Silicon Valley is full of many visionaries who build a hot new business based on a revolutionary technology, but their companies do not survive when technology or market trends change. To build a business to last, an entrepreneur, especially a "techie" type, has to realize that innovation comes not just from inventing new products, but can also just as easily come from introducing new business models and new ways to market those products.

I quickly caught up with the latest technology trends on the Internet that I had missed out on during my tunnel-vision days at Gupta Technologies. I also started to make angel investments in many startups of that era, and in 1997 one of those investments was in a little company called Keynote Systems located in San Mateo, not too far away from my home.

The company had built some interesting technology to measure the performance of Internet websites and to determine problems that might have been caused due to Internet backbone delays. But Keynote's business model was still uncertain at that time, and I quickly realized that while the company's software was not all that differentiated from many other systems management tools, we could apply an "on demand" service model to the business and make it truly valuable to up-and-coming e-commerce websites.

In effect, we would measure the real-time response time and reliability of any website on the Internet from multiple cities across the world, including of multiple competitors within the same industry or of multiple players in a product supply chain, and make this data available on a monthly subscription basis to enterprises who needed to assure themselves of their e-commerce website's technical performance and quality.

Before we knew it, Keynote was a hot Internet startup that became known as the "JD Powers of the Internet." We had more than 2,000 corporate customers across the world subscribing to our performance metrics. Though we did not know it at that time, we would also have the distinction of becoming one of the world's first "software as a service" (SaaS) businesses before the term would become popularized.

But in in the fast moving technology world, the ability of your organization to react speedily to change is just as crucial as your personal ability to anticipate the future. This was the dot-com era, and we were running a hot Internet start-up during the mother of all technology bubbles. In Silicon Valley there is a saying that goes: "if the wind blows hard enough, even turkeys can fly." While Keynote clearly had real revenues, real customers and a viable business model, I didn't fool myself into thinking I knew how our technology or business would unfold in the future. Instead I just made sure we were prepared to seize chances when they came to us.

When the chance presented itself for us to go public in August 1999, we took it way before we were showing any profits. A few months later in February 2000, when the stock market was still hot, we did a secondary offering and obtained a valuation of more than a 100x revenues. Even though the stock market bubble burst a few months later, we found ourselves in the fortunate position of having $350 million in cash on the Keynote balance sheet, and a lot of happy VCs and early investors. No question about it - I was clearly applying the lessons I learned from my Gupta Technologies days to make sure Keynote did not get left behind many of those turkeys!

The aftermath of the Internet crash of 2000 was a searing one for Silicon Valley and Keynote was no exception. Many of our customers went out of business, our revenues started to plummet precipitously, and our losses grew larger each month. While we did not have any danger of running out of cash, we did face an existential question at that time: Do we sell Keynote? Or, failing that, simply shut down the business and return the substantial amount of cash on our balance sheet to our public shareholders? Or do we try and rebuild the business, thereby risking more cash in what might be a doomed effort anyway?

There were no buyers for Keynote at that time, since at that time no one knew how far and how deep the downturn would go, and how much our revenues would decline. I was also haunted by the memory of how, during Gupta Technologies' decline, Larry Ellison had made me an offer to buy the company at what turned out later to be a pretty good price, but I did not want to sell my baby, and I had said no. After all, I had risked the future of my company many times during its first eight years as a private company and I had always managed to make it bigger and more valuable. So why would I sell to Oracle then? With the benefit of hindsight, I can say that I should have sold the company when we still had the chance.

This time around I did not want to make the same mistake again with Keynote. I had learned an important lesson by then: "Know when to hold 'em, and know when to fold 'em."

As it turned out, we did not shut down Keynote but decided to rebuild it. Over the next 12 years, we steadily regrew the company. We introduced new products for our corporate Internet customers and also expanded into the mobile monitoring and testing space through a couple of well timed acquisitions. By 2013, Keynote had grown to more than 4,000 customers, and revenues had tripled to more than $125 million with very respectable profit margins of around 20 percent.

A few months ago, we sold Keynote to Thoma Bravo, a private equity company, for around 3x revenues and a 50 percent premium above its public share price. That's a far cry from the 100x revenue valuation of our bubble days, but a pretty decent outcome for our shareholders - and yet also a good purchase for its new owners, who are looking to grow and build even more value into it.

Which brings me to the last important lesson I have learned and applied consistently at Keynote: Your company's destiny is not your destiny.

I ran Gupta Technologies like it was a mission, not a company. And even when it went public, I always thought of it as my baby. After all, it had my name on it.

With Keynote, I made sure from the beginning to recognize that my job, like any parent, was to give the company its roots and wings, and like any parent when the job was done, I would have to separate my own life from the company's life. Today, Keynote is a solid, stable company that is a leader in its space, but still has a long way to go before it will have fulfilled its potential.

During our early days of Keynote I was fond of saying, "In a trillion dollar e-economy, surely there ought to be a billion dollar company devoted to testing and measuring the online experience." I still believe that.

My own fondest wish would be, a decade or two from now, to look at a far bigger Keynote than today and say, with some nostalgia, "I had a hand in building that!"

Short Bio: Umang Gupta is a well-known Silicon valley technology visionary, entrepreneur, company founder, and public company CEO. After having spent more than 40 years helping to build the enterprise software industry, among other things being credited with writing the first business plan for Oracle in 1981, Umang is now devoting his time exclusively to the fledgling online education industry as an investor, board member and advisor.