« In Open Source Software, "The ERP Market Also Rises" | Main | Open Source Software Support for Mobile Content Coming in 2008 »

Recently, I did a feature article about the OSS movement’s effect on the business intelligence (BI) market. I came across this analysis of a May/June survey (free but registration required) fielded by Actuate, a BI vendor, on the adoption of open source software (OSS). I think it’s interesting but it might send users considering their commitment to the world of OSS terms and conditions (Ts&Cs) the wrong signals.

Before you go further, consider my biases right up front. You’ll have to read my many past blog posts on these pages or read the next two paragraphs in parenthesis. If you already know my biases relative to being an “OSS agnostic,” skip past the parenthesis.

(I believe OSS is a software development model, not a market or a business model. The development model involves code sharing across legal entities with wide-spread peer-based quality control/improvement/extension of the code. It is typically software that many people need but that does not provide anyone any particular competitive or other advantage. OSS has existed since the 1950s but the modern era of OSS dates to formal license Ts&Cs that began to appear in the 1980s and that reflect the development model. There are dozens of such licenses. Their Ts&Cs primarily involve the right to access the source code, change it and redistribute the original and/or the changes. Those Ts&Cs are different than Ts&Cs that come with software developed by programmers not sharing code across legal entities or sharing such code under restrictive intellectual property agreements. But the Ts&Cs are just different; they are not better or worse than the others’.

(More important, software licensing Ts&Cs are not a major factor in user buying decisions or overall market dynamics. Users decide which software to acquire because of its functionality. With OSS code, functionality is usually monetized the same way as non-OSS software. That is, it is monetized as a service. Such service can be delivered in the form of device/system-bundling arrangements, as subscription maintenance, as per-transaction fees and by other more creative but less prevalent means, such as giving it away in return for hoping you will read the ads that flash by you as you search the Internet. Acquiring software functionality bundled into a service or device/system or related to a service will be the predominant means that users acquire software by 2020. OSS will play a major role in that trend because its Ts&Cs concerning source availability and redistribution are more attuned to software functionality developed to be delivered as a service than the usual restrictive perpetual right to use licenses most associated with non-OSS software delivered as an executable.

(With that boring but necessary background, let me analyze the Actuate survey analysis.)

The first thing that hit me in reading the Actuate analysis is that although it relates to a demand-side survey (gauging user community plans and perceptions), it begins with another research firm’s opinion based on a supply-side (ask the vendors) methodology. I have discussed that other firm’s research here. OSS is not a discrete software market even if its usage can be measured separately—I can develop a market measurement model for any technology characteristic—but that characteristic (OSS Ts&Cs in this case) often has no meaning in understanding market dynamics.

The net-net of it is that if the OSS community hangs its hat on a supply-side metric, it suffers. By this measurement, OSS software accounts for less than 1% of the software market. If you are thinking about acquiring some software under OSS Ts&Cs, but hesitate to do so based on its small market share, dismiss market-share data.

Next, the Actuate survey only covers four countries and three industries (financial services, telecommunications and manufacturing). My comments below are based on the 390 people surveyed in North America. I didn’t see an “n” by industry so it’s hard to figure if a statistically significant number of respondents were involved from each. If you are not in one of geographic areas or industries, that does not mean that OSS is not for you.

The respondents are identified as “senior personnel” that were “invited” to participate. If the senior personnel are more IT staff than IT management, that’s a good sample base. An important question is “Who were the invited?” I have an email in to the Actuate folks and will let you know if the answer is relevant. Hopefully it is not just Actuate customers.

Analysis I have read about the survey, including in the Actuate document, highlights the reason people are not using OSS. But that appears to be based only on the answers from about 40-50 respondents so don't base any decisions on it. Of course, by my biases, everyone is most likely already using OSS (do you have an Oracle or IBM application server in your shop?) and those that aren’t will be, probably not even realizing it. By the end of the next decade, OSS Ts&Cs will just be part of the way most software is developed and monetized.

According to Actuate, the two most widely used products are Linux and Apache, each mentioned by 43.1% of respondents. Presumably that means Apache HTTP server for the reasons described in this recent feature article on ebizQ. The next most commonly used OSS distributions adopted by this group are MySQL (32.8%), Tomcat (29.5%), Mozilla (28.7%), PHP (23.3%), Eclipse (22.2%) and JBoss (15.1%). All of these statistics seem intuitively correct to me (which gives me confidence in the rest of the findings) with the exception of the prevalence of MySQL.

Counterintuitively (if that’s a word), awareness of OSS among public sector respondents in North America was lowest of the three industries. The conventional wisdom is that OSS is very popular among government users.

Posted by dennisb in OSS Business Issue |Digg This|Add to del.icio.us

Trackback Pings

TrackBack URL for this entry:
http://www.ebizq.net/mt/mt-tb.cgi/2872