The year is 1999, Google is still a word that only math majors are familiar with. A startup called Kozmo will hand deliver a snickers bar and a DVD rental to your front door in 15 minutes — for free, and no tips allowed. A company called MP3.com will give you a free set of high-end computer stereo speakers complete with sub woofer and 50 high quality MP3 files — just for signing up on their website.
It was into this environment that I launched Extentech — a bootstrapped boutique web development and web hosting, and consulting firm. Initially our plans for breakout success revolved around creating a turn-key website appliance to allow SMBs to “get online” and establish a web presence. In fact you can still see the original Extentech site here.
This was a time of massive hype and minimal substance — and I was going to try to differentiate on quality and actually building software systems, not just web-based brochure ware.
Some of our many expenses included the need to build and host servers ourselves and on behalf of our customers. There was good crossover work and a steady hosting income to supplement revenue lulls between projects, but the ability to subsidize some product development without having to raise capital was enticing.
What we didn’t have at the time was a product, and of course having a software product and selling a license to use it and support is a much more scalable business model than selling your limited time as a consultant.
After a couple of years of consulting, we were eager to build something more lasting and that opportunity came in the form of a Java Spreadsheet tool, ExtenXLS. We basically took on the hardest project we could find, and tackled a customer need that nobody else wanted to mess with — reverse engineering the Excel file format.
We figured that everyone uses spreadsheets in business, but not everyone was running Windows — especially in the Enterprise. So ExtenXLS was born as a way to add Spreadsheet functionality to Java applications. This turned out to be a fine idea, and one that has sustained our business for over 10 years.
After a decade of fighting with some of the trickiest coding issues around (try grokking the BIFF8 spec for a mind numbing sleep aid) we developed the most robust Excel-compatible spreadsheet toolkit for Java.
Extentech was a business that evolved from circumstance and customer needs. We did not start out with a plan to develop a spreadsheet SDK, nor did the AJAX technology upon which we based Sheetster our online version even exist at the time the company was founded. We basically pivoted our way to success, tweaking our business model and positioning based upon the changing marketplace.
When Google muscled into our space with Google Docs in 2006, we were forced to rethink our plan for a consumer-based online spreadsheet, instead rebuilding Sheetster from the ground-up as a development tool. The positioning of Sheetster as an embeddable, Open Source alternative to Google Docs and Excel was forced upon us, but in fact resulted in a more sustainable differentiation.
By the time Infoteria purchased Extentech in May 2012, we had many enterprise customers, and a sizable demand for our products.
So, in the 14 year journey that was Extentech, we faced innumerable challenges, built our own tools that predated the existence of much of the available platforms today. For example, when we first began building web applications, there were very few JSP-compatible application servers available — so we built our own. In fact, over the years, we built and dog-fooded our own security frameworks, blogging tools, KB, CMS, and eCommerce systems, app servers, and reporting tools. I even hand-coded a Java-based SMTP server in the early days. With the availability of SaaS offerings, and open source tools, the vast amount of platform code and tooling that we built ourselves would be totally unneeded today.
With the tools and services available today for free and/or marginal cost, without a doubt we could rebuild everything we did at Extentech over the years in a fraction of the time. Today, developers with one tenth of the experience can build software 10 times faster than 10 years ago. Business models like Kozmo’s have come and gone, and the inefficient have fallen by the wayside. Had our product stack had a lower barrier to entry and lower value to the customer, then there is no doubt we could have shared that fate.
We were lucky in that we picked a product to develop that had a huge demand, was well differentiated, and carried a high barrier to entry. These ingredients to success are still vital and in fact even more so now that the technological barriers have fallen across the board. Moving products up the food chain from SDKs and platforms to vertical applications that leverage best practices and industry expertise are the winners in today’s market of ideas.