Ask ten different companies about their cost-savings experiences when switching to open source, and you get ten different breakdowns according to ten different timelines and metrics. Shouldn't we have something a little more consistent than that?
This line of thinking started after I read two articles, "The open source sea change and the Taylor graph" (Dana Blankenhorn @ ZDNet) and "Specsavers sees clear benefits in open source" (ZDNet.co.uk). The former questioned the conceit that long-term costs drop dramatically after a switch to open source; the latter is a case study that asserts this is precisely what happens ... although, it's a single case study involving a single company.
That's long been the problem with these sorts of studies -- there are so many of them, and done in so many different ways, that picking through them can be like cleaning out the Augean stables with a whisk broom. Not much of a help to open source if the very data that supports their premises is only available in such a far-flung fashion.
Here's my thought: Instead of a hodgepodge of studies conducted by different groups, maybe what's needed is a single, publicly-maintained questionnaire or other data-collection instrument that those who've switched to open source can fill out. The resulting data can then be aggregated across multiple companies, industries, and public sectors -- everything on one page.
This way, success stories like Specsavers could be stacked up against the narratives provided by any number of other companies, good or bad. In short, it'll be a financial / business-practices version of things like the Debian/Ubuntu package usage survey.
I haven't forgotten the larger question of who maintains and produces such an instrument. This part isn't as important, as long as the people doing it are able to keep their work clearly documented and can afford to keep it up year after year. The hardest part is actually not the upkeep -- storage and bandwidth are cheap -- but drafting the instrument itself, something best done by people who have existing expertise in gathering such data.
Open source itself has evolved aggressively. Now let's see the tools with which we assess and monitor the use of open source follow suit.
InformationWeek Analytics has published an independent analysis of IT governance models and metrics. Download the report here (registration required).