Why Companies Really Turn to Open Source Software

I blog primarily to set the information technology (IT) investment-research record straight in response to out-of-bounds or misleading marketing claims from leading software suppliers. I consider it penance for the 20 years I spent making out-of-bounds or misleading marketing claims for two of the leading software suppliers back in the day.

Most such claims are in press releases or SEC filings but occasionally I run into press articles that "round up" a half dozen such marketing misdirections from various marketing departments in one place. The December 1 BusinessWeek article called "Cost-Conscious Companies Turn to Open-Source Software" is such an article. And it's the second such multi-marketing-oriented misdirection article I've seen this week. I think it's because editors are trying to link IT to the economic downturn.

To be clear, I am a big fan of the open source culture. But this BusinessWeek article and similar articles are so full of misleading points that they put open source in a no-win situation: Open source cannot deliver on the hype that suppliers and PR writers are weaving around it and as a result the culture will suffer in the long run. This hype also misleads investors although that trend has been seriously tamped down since the 2007 peak when Yahoo (YHOO) paid a zillion times revenue for Zimbra.

What's really surprising is that this article is a BusinessWeek deliverable. It sure doesn't meet the McGraw-Hill standard I had to follow at Datapro. The author is repeating urban legends, which is not unusual. But the author does not even pretend to substantiate them, by finding an obscure source to quote. Despite what the "Cost-Conscious..." article says or implies, the facts about open source software are as follows:

• Most popular enterprise-level open source software is written by programmers who work for the company that hopes to benefit commercially. As a result, they "...work for one company" (although I am not sure what that factoid was meant to prove anyways). Thousands of non-commercial, non-enterprise-level open source projects are written by academics and hobbyists and that's what's great about the culture. (Of course they also "work for one company" or organization.) Open source is a culture that is as old as the IT industry and not a new trend in the last few years.
• Research does not indicate that the total cost of ownership (TCO) for open source software is any different than for non-open source software. Nor would I expect it to be because software licenses are such a small part of the TCO equation (refer again to the linked blog post about the misleading claims by Ingres' CEO on open source and the economy).
• Particularly, despite the bald statement in the article, there is no indication that users "pay less for a subscription to an open-source product than for its proprietary counterpart." In fact, one of the basic premises of the core open source community is that the software is "free as in speech," not "free as in beer."
• Of course it is hard to find open-source vs. non-open-source counterparts to compare. That is why open source exists. Typically open source software is developed where a given functionality has reached commodity status and there is no particular competitive advantage between one type or another. Therefore systems and software suppliers do not want to invest in separate development and instead they choose some open source functionality to embed in a proprietary product and/or to provide their customers directly as a convenience.
• Where there are price differences between two roughly comparable software products/services, one available with open source terms and conditions (Ts&Cs) and the other not, it relates to the fact that, as the BusinessWeek article says, users "will get a lot less hand-holding than they may be used to." If you don't need the hand-holding and the specific open source functionality is what you need, open source is a good choice.
• Typically however IT users want the hand-holding, even for commodity software. This is the reason for Red Hat's (RHT) success, and this is the reason that most enterprises using the Apache HTTP web server (probably the most popular piece of open source software), use a version embedded in a so-called non-open-source suppliers' products [e.g., built into IBM Websphere].
• By the way, for investment purposes, the distinction between non-open-source supplier and open-source supplier is a canard. All leading software suppliers, including Microsoft (MSFT), distribute software using open source Ts&Cs and almost always have.
• The BusnessWeek article says "Linux as an operating system for the desktop is much less established in enterprise use, with 39 percent of respondents in the Gartner survey currently using it and another 22 percent expecting to use it within the next year." While I am sure the Gartner numbers are accurate, the reality is that Microsoft Windows dominates the desktop followed by Apple's (AAPL) operating software. Linux's share of desktop usage is in low single digits. That's not a criticism of Linux; Linux is UNIX software and UNIX was never intended to be desktop software. In fact, UNIX predates the concept of desktop software.
• There is no research that open source software is more reliable than non open source software, or more secure (not mentioned in the article but another popular canard). Open source from a business perspective (as opposed to a cultural perspective) is basically a set of Ts&Cs and has nothing to do with things like reliability, security or any particular software functionality.
• There are dozens of open source licenses with varying types of restrictions and Ts&Cs. "License management" is specifically not an advantage, despite what the BusinessWeek article says, confusing acquisition activity with license management. In fact license management is one of open source's drawbacks (see below for its advantages).
• Most software--no matter the type--is test driven first before it is purchased or a subscription is obtained. There is nothing different about open source software when it comes to "try it, you'll like it" marketing.
• IBM is probably the largest provider of open source software services, not Red Hat. Oracle (ORCL) is coming on strong in stealth mode, actually distributing a Red Hat clone (making money, as Red Hat does, by servicing it) but also distributing Apache software and its own open source database product. Microsoft may also be one of the leading distributors based on an arrangement with Novell (NOVL). When the article says "About two years ago, Office Depot began installing the Linux operating system by Novell on its servers, moving away from proprietary software for a range of hardware including IBM mainframes and Sun (JAVA) servers," that could very well be because of an arrangement with Microsoft. Of course Sun also now distributes its software with open source Ts&Cs.

In the real world, as opposed to the PR world in which this article seems to spin, companies turn to an open source product primarily because it does (or they perceive it does) something better than the other software they looked at. Often companies take it without hand-holding as described above but most often they want support and service, making it no different from a market-research (and therefore IT investment research) perspective than software with other Ts&Cs.

Second, companies choose open source for the benefit of having perpetual access to the source code itself. That's a good reason, but is much less a market dynamic than a question of functionality. Most companies would not want the hassle of having to maintain code if the creator of the code no longer existed. It is more important to use increasingly scarce resources, even IT resources, to do things that make the company's product or service different in its own market.

A very small number of companies do choose open source simply because of the way it is developed; that is, for the culture. Similarly, some people will only drink French wine or wear Italian shoes.

Comments

[David Link:]

Dennis,

Thank you for taking the time to expose one of the distinguishing characteristics of a segement of our current generation of technology journalists. They don't dig for the facts to verify and build credible evidence for their ascertions. This pattern of young, less than experienced journalists continues to distort a full perspective of the topics they choose to cover. I enjoyed not only your perspective, but the fact that you revealed a bit of your past professional experience in which you delivered the context of your opinions. We need more refreshing articles and resulting conversations just like the one you published today!

Jan 19 09:25 AM

[donzelion:]

Emphatic concurrence here: "Research does not indicate that the total cost of ownership (TCO) for open source software is any different than for non-open source software." Open-source software does seem to change the structure of costs - money spent on programmer/techies rather than royalties. That can save money, but can also cost a great deal more (particularly in terms of lost productivity).

But I have to say this much: open source or closed source, I've never heard of IT professionals selling financial instruments that obliterated the buyer.

Jan 19 12:11 PM

[Jonathan Scott:]

I enjoyed your articled.

Did you say people turn to open source because their immediate goal is a common goal that can be accomplished with a commodity-like solution, and some people turn to open source like some people buy french wine and italian shoes?

Jan 19 03:44 PM

[Zaskoda:]

"Second, companies choose open source for the benefit of having perpetual access to the source code itself."

I like to call it "intellectual sustainability". I typically reference Web browsers as the classic story example. A technology was birthed. Then it was improved on competition. A commercial company (IE) managed to dominate the market using methods other than (just?) a superior product. Once it dominated the market, innovation came to a halt.

It wasn't until an open source project (Firefox) came along that innovation ramped up again. So long as the tech was locked down by a single commercial company, we saw very little in terms of real innovation.

I've noticed this pattern in other technology and would love to see someone do a real investigative report on the topic. At the end of the day, I choose to deploy open solutions largely for ethical reasons.

The interesting connection is that the "ethical" reasons are very directly logistical. Ethically, I can say that it's better for our culture, as a whole, to share knowledge openly. However, logistically, I can also say that having the source code in hand empowers me - individually - to make any changes I need to - and make them when I need to.

It's hard for me to imagine seeing things any other way.

Jan 19 05:04 PM

[MarcL:]

appreciate your perspective, and agree with a lot of what you point out. however, from a business perspective there are some additional advantages offered by the open source format.

* it removes the up-front capital requirement (get's rid of the initial license expense)
-- this is an important consideration particularly in the current recession and credit market. removing capex can be a substantial business benefit and can enable a project to move forward which otherwise would have gone on hold because of the up-front license costs of a 'non-open-source' system.

* enables a company to prove out the software system before committing
-- most 'non-open-source' systems require that you pay for the licenses before you know if the product will work for your specific requirements. even if a 'non-open-source' provider offers a 'trial' version this is impractical for an enterprise system of any size / scope (it's difficult and costly to implement, integrate, migrate data, etc onto a trial). the open source format eliminates the risk of paying for something before your company knows it will do the job.

there are other business benefits as well that the oss format can provide over the conventional 'non-open-source' structures such as a predictable fixed cost structure and the ability to avoid lock-in.

it appears to me that oss represents a growing new segment in the mature enterprise software market. i work for a company called Aras that offers an open source plm software system, and we believe that the oss format provides companies with an alternative structure for doing business.

Jan 19 05:48 PM

[Josh Thompson:]

"So long as the tech [browser] was locked down by a single commercial company, we saw very little in terms of real innovation."

The scenario is similar to record labels previously controlling what was released compared to initially pirating and eventually the free for all on youtube, blog sites, itunes, etc.
It does confirms the opening up of a new distribution channel is democratizing to innovation. This process is not unique to software. 15+ years ago I was told that the overnight delivery business model destroyed the concept of a distributorship because products could be shipped directly from the manufacturer.
An unanswered question is why can't a large for-profit company recreate the innovation of open source? The flaw of Open Source is at its core it is saying: "can't we all just work together...who needs self interest?" When MSFT creates the "feeling" of an open source community and a marketplace for that community with Azure developers it will all go back to normal.

Jan 20 02:22 AM

[jack dee:]

Open source has made more money for those writing about it then anything else.

Jan 20 05:19 PM

[BeachRider:]

Open Source not only has a Capital Expense effect, but also an Operating Expense effect. The salary, benefits, envrionment, HR-overhead are clearly the costs that offset lowered-capital in Open Source scenarios. Software maintenance and forced, cyclical upgrade/expansion costs are the alternative for non-Open Source situations.

Probably the biggest risk of Open Source is program-development risk. It is low if the GNU toolkits basically meet the need, but it escalates when program-development becomes more complex. IRS spent Billions in a partial-open-source environment that was developed over 8 years. They ended up with a pile of mush and tremendous embarrassment. THAT is the greatest risk.

The open source, itself, wasn't the risk, but the responsibility to make it function as a system is.

Jan 22 06:16 PM

[bigmoney:]

>> Research does not indicate that the total cost of ownership (TCO) for open source software is any different than for non-open source software

this depends on the usage. A company like Google could not scale out the way it did if it had buy a Microsoft Windows license node-by-node, CPU by CPU.

Jan 23 11:52 PM

[sammy spade:]

> * it removes the up-front capital requirement (get's rid of the initial
> license expense)

Apparently you've never looked at the pricing for RHEL. And, oh, if you stop paying the subscription, you stop getting security updates. Even MS doesn't extort you like that.

> * enables a company to prove out the software system before committing
>
> -- most 'non-open-source' systems require that you pay for the licenses
> before you know if the product will work for your specific requirements.
> even if a 'non-open-source' provider offers a 'trial' version this
> is impractical for an enterprise system of any size / scope (it's
> difficult and costly to implement, integrate, migrate data, etc onto
> a trial).

Again, not unique to open source. (your obviously only familiar with oss, and assume that all the proprietary guys suck) Sure, you can use CentOS to try it out before you fork out for RHEL. Likewise, you can use MSDN to build and test your software on MS before you shell out for the server licenses. And the big difference is...?

> there are other business benefits as well that the oss format can
> provide over the conventional 'non-open-source' structures such as
> a predictable fixed cost structure and the ability to avoid lock-in.

Sure, if RHEL doesn't pan out, you're not "locked in" and can just jump back to Solaris? Oh, that's not what you were referring too? I suppose you can port to SUSE, or Debian, but it's not entirely trivial.

The lock-in thing is a bit of a myth. Lots of business software can run on Windows, Linux, or UNIX. SAP, Oracle, even Apache stuff.

>
>
> it appears to me that oss represents a growing new segment in the
> mature enterprise software market. i work for a company called Aras
> that offers an open source plm software system, and we believe that
> the oss format provides companies with an alternative structure for
> doing business.

I bet you do.

Feb 22 11:42 PM

[sammy spade:]

> Software maintenance and forced, cyclical upgrade/expansion costs
> are the alternative for non-Open Source situations.

Sort of. If you look at Red Hat support, support starts to drop of after 4 years. MS still supports Windows 2000 and XP (granted, XP is a "new" OS for netbooks, but I digress). Seems like your forced to upgrade either way, unless you just want to run the same app, on the same OS, without ever applying patches. In that case, you're not forced to do anything.

Feb 22 11:51 PM