Linux Today: Linux News On Internet Time.

Andre Boisvert: The Man Behind the Open Source Curtain

| | Comments (0)

By Mark Hinkle

Most open source luminaries are known for their code, their successful startup successes or even their outspokenness. Andre Boisvert comes to open source from a different angle. Having worked for two billionaire programmers, Larry Ellison and Jim Goodnight, Andre's transition from proprietary software to open source software has been an interesting journey.

Smashing PumpkinsAndre started out his career at IBM where he spent 13 years. At Big Blue he was fast tracked through various positions in sales, marketing and R&D as part of their executive program. He then left for a turn around at Cognos (now owned by IBM). He's been the President and COO of the world's largest private software company, SAS Institute Inc. and has been the SVP of Marketing at Oracle. After working for some of the largest proprietary software companies, Andre now works primarily advising open source startups using his experience in order to help them better compete with some of his former employers. Though he keeps a relatively low profile in the open source community though he's definitely a mover and a shaker.

The Counter-Intuitiveness of Open Source

In 2002, after joining the board of VA Linux (now SourceForge) Andre got exposed to the many thousands of open source projects on SourceForge. At first a lot of things in open source software were counter-intuitive to Andre. For example, how could volunteers and distributed developers outperform a dedicated development staff? For example, at SAS he had in excess of 1,000 developers housed on SAS campus, which felt like an extension of academia. With numerous accolades as one of the best places in America to work, the employee turnover at company was virtually non-existent. As a privately held company with a virtually unlimited budget there was no pressure to partake in the unnatural acts required of publicly traded companies to appease investors.

The problem with this was with very little fresh blood flowing in there was less innovation. Beyond that even though the developers within the company reviewed each other's code there was a downside. If the guy in the next cube evaluates your work and dings it, he has to confront you every day at work, at your kids Little League practice, the company cafeteria, etc. This makes it less likely for someone to really challenge the product. In the OSS community while some developers are sleeping others on the other side of the globe is scrutinizing your code.

To illustrate his point Andre notes the phenomenon of the typewriter tough guy:

We are all recipients of flaming emails at least once in our life's and typically that email comes from someone we never met or at least someone that isn't standing in front of us." The same thing happens in a worldwide open source code review. It drives programmers to write better code because of the public nature of OSS. On the other hand if you are a good programmer and you are willing to take criticism you are able to partake in a community that can offer genuine help in near limitless supply.

Open Source Software (OSS) is very democratic. Individuals can influence the direction of software. While many talk about collaborative software contributions are generally in the form of plug-ins and add-ons but collectively the large number of users give input on the direction of the core technology. This is a huge advantage over proprietary software development where asking existing customers about your product doesn't really help because they already bought it. Plus they aren't going to give you a beat down after a free dinner and a round of golf.

In short what Andre learned about open source was that:

  • Code quality was better through a more transparent distributed development process.
  • The cycle of innovation is accelerated by a relatively larger group of users with diverse needs (carrot) and the constant threat of being forked (stick).
  • The barrier to deploy open source software is lower since there is typically no perpetual license up front, just an "optional" support contract.
  • Full code availability makes it easier to integrate with other systems compared to that of proprietary software for which their respective vendors offer poorly documented APIs. Additionally, open source developers tend to adhere to open standards.

In late 2007 Goldman Sachs cut their estimates on many software companies. Their sentiment was consistent with the value being provided by many open source software companies.

Concurrent with our change in coverage view, we are reducing the estimates of most of our covered companies, focusing on pure-plays that could be harmed as customers seek to purchase 'good enough' substitutes from larger vendors, as well as vendors who sell 'big ticket' items that could be delayed in a slower spending environment," Goldman Sachs analysts wrote in the report.

According to Boisvert they also noted that, publicly traded software companies are spending between 83% to 100% of new license revenues on sales & marketing activities. Customers not only pay for their own golf outings but all the outings of all other prospects that didn't pan out.

What's been a challenge for a lot of open source vendors in their sales process is that there is little room for free pre-sales support, there's no free dinner at Morton's steak house. The economics are different because there is no perpetual license for the vendor to recoup these sales and marketing expenses. Under the open source model, customers have come to realize that they need to pay for pre-sales activities such as "proof of concept pilots" and that golf outings are not included.

The new model typically starts with a customer/prospect downloading the software and then reviewing it on their schedule and not that of the vendor. Customers can check out the programs in detail and hear unbiased views by going on project forums as opposed to listening to a yapping sales guy shoving glossy sales brochures under your nose and making promises he knows are not likely to be kept.

Building Open Source Companies

Pentaho: Open Source BIAndre was anxious to put the lessons learned at VA software to use and in October 2004, he co-founded an open source business intelligence company , Pentaho. Since the company had no install base, they had the luxury of the creating the next generation Business Intelligence (BI) platform. This was a feat the "popriatary" industry leaders could have never achieved, as migration issues from massive legacy customer bases would have placed too many limitations on the product.

The Pentaho crew knew that having a vibrant community was key to their success and unlike many companies that labor over building communities over a long period of time, they acquired the IP rights and hired the key developers of three projects: Jfree Reports, Kettle, Mondrian. On their own they were all good pieces of software but combined into a BI platform, the sum was much greater than the parts. The communities had similar interests and they came together rather quickly offering even more synergy.

Compiere: Open Source ERPAndre's appetite for open source didn't stop at Pentaho, he was soon the first outside investor in Compiere. Compiere founder Jorg Janke, a former Oracle colleague, had spent two years building an ERP platform from scratch. Jorg understood the problem with the then current proprietary ERP systems which drove him to build the next generation ERP platform.

Andre shared that view with Jorg since as head of Oracle's worldwide marketing function he remembers Larry Ellison would evaluate feature requests by saying, "If we had the features that this specific customer is asking for, how big would the check be?" Depending on the size of the check, Oracle would build out features which in many cases was only applicable to one customer. Being the marketing guy Andre would spin the features as designed for specific "vertical markets" but at the end of the day it was a feature used only by a single customer. Other customers ended up paying for the additional overhead to run this code (MIPS, storage, etc) and to add insult to injury, paid for it every year on their maintenance bill.

Jorg knew that every customer had a common need, financials. Jorg built a financials package that supported multi-currencies, multi-nationals, multi-divisions, etc. His data dictionary approach allowed customers to only generate the code the user needed. Unlike many open source projects there was not one outside code contributor and his partner Kathy Pink did all the QA and documentation. Janke could have released Compiere as proprietary software but after a long conversation with JBoss' Marc Fleury, Jorg decided to release the code to the OSS community. In parallel Andre and Larry Augustin both then sitting on the board of VA Software saw the market opportunity for the software and subsequently helped fund the company.

While Compiere's product was entirely built internally, Pentaho chose to combine both internally developed code and acquiring a handful of complimentary open source projects. This allowed Pentaho to deliver an end to end BI platform which includes things like ETL, Analytics, and reporting that formed. Both open source companies are the leaders in their respective spaces but they each took a very different approach in getting there.

Having been an investor and chairman in a pure applications company like Compiere, and a company that straddled application and infrastructure, like Pentaho, Boisvert, driven by curiosity and a desire to spread his investment across the software spectrum looked at systems management. He soon became an investor and the chairman of the board of an open source monitoring vendor, Zenoss.

Zenoss: Open Source IT Monitoring

Andre was intrigued by how much open source software was being used in larger enterprises, especially by the open source companies that he was associated with, therefore Andre joined the board of Palamida. Palamida is a company that helps mitigate IP risk in open source software. Being on the board of Palamida, Andre gained valuable insight into which applications are being used in the enterprise and the required attributes for OSS offerings to be adopted in the enterprise.

Playing Both Sides of the Fence

Boisvert is also frequently asked to speak to buy side analysts at large investment banks. Not to sell them open source stock, or at least anytime soon. He actually shows them how to hedge their bets on proprietary software companies who might be losing ground to open source challengers.

Similarly Andre still participates on the boards of proprietary software companies like UBmatrix Inc. the creators of XBRL (Extensible Business Reporting Language), Odessey Technologies Financials Inc. a Swiss-based leader in wealth management software and more recently joining as the chairman of the board of Infobright, a proprietary analytic data warehouse provider. However, one reason he's attractive to these companies beyond his proprietary experience is to advise them on the potential use of open source software within their own respective software stack and therefore lowering their R&D costs.

Andre's thoughts on open source can be summed up with this:

As the US and Europe have an economic slowdown, IT is the weapon of choice for improved efficiency. Open source is the needed ammunition which can be had at disruptive price points, not just incremental. Additionally, the economies of the BRIC [Brazil, Russia, India, China] are growing at a staggering rate and they have been raised on being cost conscience, a perfect play for open source. As such, this is the best time to be in the open source business.

Creative Commons License photo credit: Brenda Annerl

[Disclosure: Mr. Boisvert is the Chairman of the Board at Zenoss where I am on the executive team]

Comment and Contribute

    (Maximum characters: 4000). You have 4000 characters left.