Halloween VII: Survey Says

The document reproduced below was presented at a Microsoft internal Linux Strategic Review held at the Microsoft offices in Berlin during Sept. 2002. I received it on 5 November 2002.

What We Can Learn

Here's a summary of the tactical advice for open-source advocates that I think we can glean from this memo:

  • The messages and tactics the open-source community has developed over the last five years are working well. Our memes about security, TCO, and competitive impact have achieved deep penetration in Microsoft's survey population. Abstract arguments about intellectual property rights, on the other hand, have served Microsoft just as poorly as they have served us.

  • Microsoft's FUD attacks on open source have not only failed, they have backfired strongly enough to show up in Microsoft's own market research as a problem. This means we don't need to put a lot of energy into anti-FUD defending the open-source way of doing things. Indications are we've won that battle; effort should now go elsewhere.

  • We need to keep Microsoft's feet to the fire on the TCO issue. Their figures indicate that we're winning that battle (no surprise, especially not after the XP licensing changes). If the memo recommendations are followed, Microsoft will attempt to reverse this with all the money and marketing clout it can muster.

  • Familiarity with open source makes respondents less vulnerable to Microsoft's ‘shared source’ scam. The higher respondents scored on familiarity with open source, the less likely they were to judge that shared source offers the same benefits. We need to keep hammering on the difference between source that you can see only after signing a Microsoft NDA or non-competition agreement and source that anyone can examine, modify, and redistribute. Emphasizing the poison-pill problem is indicated.

  • Internationally, a distaste for being dependent on U.S. technology companies in general (and Microsoft in particular) is exploitable. Microsoft perceives serious problems with this, as well it should.

  • High approval has not yet translated into wide deployments. More managers like Linux in theory than routinely use it in practice. This suggests that many are either waiting to see results from large path-breaker deployments by others or are hampered by organizational inertia.

  • The risk that Microsoft will go on a patent-lawsuit rampage, designed more to scare potential open-source users than to actually shut down developers, is substantial. The language about “concrete actions” in relation to IPR has the same ominous feel that the talk of "de-commoditizing protocols" did in Halloween I and II.

  • The term ‘free software’ isn't mentioned once, not even as an exploitable weakness. This contrasts strongly with the original Halloween Memoranda. I'm not sure what this means, but one strong possibility is that the term has simply fallen out of use both at Microsoft and in their survey population.

The overall tone of the memorandum is very defensive. Not quite panicky, but the researchers are not able to name any argument with the open-source community that their own figures show them to be winning.

In fact, their figures indicate that we are winning. It looks like all we have to do is stay the course.

Reading The Memo

Some helpful vocabulary, deduced partly from usage in the document itself and partly from looking at Microsoft documents that have leaked onto the Web.

BDM

Business Development Manager: fancy term for IT salespeople and IT sales managers.

Developer

A non-manager. In this survey, a disjoint category from “IT Pro‘. Therefore, it probably means anybody with a technical clue.

IT Pro

IT middle manager, what we'd call a PHB (Pointy-Haired Boss).

Issue Elites

Seems to refer to both policymakers in education/government and (though less certainly) influence leaders among strategic-level corporate executives.

Bad markup generated by Microsoft's broken-as-usual HTML tools has been corrected. Sections in this color are portions I think particularly noteworthy. My comments are in this color.


Research E-Bulletin: Attitudes Towards Shared Source and Open Source Research Study

Due to the sensitive nature of this information, please forward with discretion only to those people who can clearly gain value from it.. For those members of the Linux Strategic Review Core and Virtual Teams, this information is for background use/understanding during the Linux Strategic Review.

{Well. I guess I qualify as a person who "can clearly gain value". Nice to know I'm legally in the clear!}

Executive Summary

This mail provides a detailed summary of the results of the Attitudes Towards Shared Source & Open Source Research Project managed by Kathryn Marsman and directed by David Kaefer and Jason Matusow. The Shared Source project was developed to provide a greater understanding of how key audiences perceive Open Source, Linux, Shared Source, and the GPL and which messages will be effective with each audience. The survey was fielded in the U.S., Brazil, France, Germany, Sweden, & Japan with developers, IT and non-IT BDMs, IT Pros and Issue Elites. Please note that save for the U.S., the individual country and audience sample sizes are extremely small. The survey questionnaire and samples were developed collaboratively by Redmond, the subsidiaries and the survey vendor. All data collection utilized a telephone-based interviewing process. The study fielded between late-July and September 2001. The detailed summary below drills into OSS and Linux familiarity and favorability, those reasons people give for being supportive of OSS and Linux, Shared Source familiarity and favorability, and OSS, Linux and Shared Source messaging. Key takeaways follow.

  • Familiarity and favorability for OSS and Linux was high across geographies & audiences. Eighty-one percent (81%) of respondents Worldwide said they were at least 'somewhat' familiar with OSS; 77% of respondents Worldwide said they were at least 'somewhat' familiar with Linux. Worldwide 78% of OSS familiar respondents said they had a favorable impression of OSS; Linux favorability among the Linux familiar was 86%.

    {86% Linux favorability in a Microsoft survey!}

  • While respondents cited OSS's 'low Total Cost of Ownership (TCO)' as one of the best reasons to support OSS, an 'alternative to Microsoft' did not lag far behind. A plurality (40%) of all respondents felt that a low TCO was the best reason to support OSS. One-third of all respondents cited 'an alternative to Microsoft' as one of the best reasons to support OSS.

  • Though familiarity with Microsoft's Shared Source initiative is low, the reaction to Shared Source was positive. Thirty-nine percent (39%) of respondents said they had heard 'something' about Microsoft's Shared Source initiative, while 60% said they had heard 'very little' (35%) or 'nothing' (25%). When read a brief description of Microsoft's Shared Source initiative, the reaction is more positive (47%) than negative (15%). The other third of respondents said they were ‘neutral’ towards Shared Source.

    {So, open source was viewed more favorably than ‘shared source’ both before and after Microsoft's pitch.}

  • Messages that criticize OSS, Linux, & the GPL are NOT effective. Messaging that discusses possible Linux patent violations, pings the OSS development process for lacking accountability, attempts to call out the 'viral' aspect of the GPL, and the like are only marginally effective in driving unfavorable opinions around OSS, Linux, and the GPL, and in some cases backfire. On the other hand ‘positive’ OSS, Linux, and GPL messages are very effective - both across geographies and audiences.

    {I've suspected for a while that the anti-Linux, anti-GPL FUD campaign was actually rebounding on Microsoft. This seems to confirm it.}

  • Shared Source messages that offer transparent benefit(s) and are audience specific ARE effective. Partner audiences (IT BDMs and developers) are encouraged by messages that indicate that Shared Source will make it easier to build applications based on APIs and that Shared Source will build the developer community. Customer audiences (IT Pros and Non-IT BDMS) respond best to improvements in the feedback process, and being able to perform security checks. Issue elites respond extremely positively to the potential for increased education access to the source code.

    {In other words, Microsoft's own customers have bought the security argument against closed source.}

Closing, those who are familiar with OSS and Linux are favorably predisposed towards them. Linking this work with other on-point research, we can assume that in the majority of cases this reported 'favorability' is more emotional than it is rational. Given this context, we should not expect rational arguments focused on undermining support for OSS, Linux and the GPL to perform well. In the short term, then, Microsoft should avoid criticizing OSS and Linux directly, continue to develop and aim to eventually win the TCO argument, and focus on delivering positive Shared Source messages that contain transparent, audience specific proof points.

{So they're going to try to beat us up on total cost of ownership! Well, it was all pretty sane up to that point; it's nice to know our opponents are still smoking crack about some key issues. Or have they simply escaped noticing that the worst customer revolt they're dealing with is over the TCO implications of the new XP licensing scheme?}

Detailed Summary

Familiarity & Favorability of Open Source Software (OSS) & Linux

  • Open Source and Linux are well know(sic) and well regarded in these communities. Overall, a majority of respondents are familiar with OSS (81%) and Linux (77%). A solid core are ‘very’ or ‘fairly’ familiar with OSS (51%) and Linux (41%). Among those familiar, strong majorities have a favorable view of OSS (78%) and Linux (86%). Although favorability is high, the majority of respondents rated Open Source and Linux as “mostly” favorable as opposed to “very favorable” indicating that their favorability was not strong.

  • Familiarity and favorability for OSS and Linux was high across geographies & audiences.

    • Overall Familiarity, while high among all countries, was highest among the Japanese. Japanese respondents were the most familiar with both OSS (88%) and Linux (87%) with nearly 9 out of 10 respondents at least ‘somewhat’ familiar with each. Note though that the large majority of Japanese respondents said they were only ‘somewhat’ familiar with both OSS (81%) and Linux (77%). The reported degree, then, of their familiarity was lower than the other countries surveyed.

    • Among those aware, favorability was highest among the Germans, French, and Brazilians. Favorability was high for both OSS and Linux among the Germans (86% and 93%, respectively), Brazilians (85% and 90%), and the French (87% and 89%).

    • Not surprisingly, familiarity of OSS and Linux among the individual audiences was highest among Developers. A high percentage of Developers were familiar with both Open Source (87%) and Linux (84%).

    • Among those aware, favorability was highest among the Issue Elites. Favorability among Issue Elites was high for OSS (86%) and very high for Linux (95%).

Support for Open Source Software & Linux

  • Overall respondents felt the most compelling reason to support OSS was that it ‘Offers a low total cost of ownership (TCO)’. Forty percent (40%) of all respondents felt that a low TCO was the best reason to support OSS; however an ‘alternative to Microsoft’ was a strong second with 34% overall.

    • French respondents exhibited a strong anti-Microsoft sentiment as sixty-one percent (61%) stated that ‘an alternative to Microsoft’ was the most compelling reason to support OSS. This sentiment was echoed to a lesser extent among the Germans (37%) and Swedes (35%).

    • Among the individual audiences, Elites selected ‘an alternative to Microsoft’ as their primary reason for supporting OSS (46%).

  • French, German, and Brazilian respondents were the most convinced that Linux offered a low TCO. Sixty percent (60%) of French, 57% of German, and 53% of Brazilian respondents believe that a Linux solution offers a lower TCO than proprietary software.

  • Although familiarity and favorability were strong for Linux, overall only a quarter of IT respondents were interested in broadly deploying “Linux in your business.”

    • Only 24% of IT respondents Worldwide were interested in broadly deploying Linux in their business. But, respondents in Germany & Japan do pose an immediate concern. Half of all German IT respondents (50%) and nearly forty percent (37%) of Japanese IT respondents were interested in broadly deploying Linux within their business.

    • When split by audience, only IT BDMs showed a strong interest in broadly deploying Linux, with about one-third (33%) stating interest. This was distantly followed (26%) by IT Pros.

Familiarity & Favorability of Microsoft’s Shared Source Approach

  • While US respondents were the most likely to have heard about Shared Source (91%) — followed by the Japanese (86%) and the Swedes (81%) — most respondents had heard only ‘very little’ about the initiative. The French were the least likely to have heard anything about Shared Source with only 63% saying they have heard “nothing” at all.

    • IT Pros and Developers were the most likely to have heard something about shared source (79% each). However, the large majority of respondents reported hearing “just something” (24%) or “very little” (35%) about shared source. Twenty-five percent (25%) of all respondents -36% of Issue Elites - had heard ‘nothing at all’ about Shared Source.

  • The Reaction to Shared Source is more positive than negative. When read a brief description of Microsoft’s Shared Source Initiative, while the reaction is not as overwhelmingly positive as the reaction to OSS, 47% say that having heard this description their reaction is at least ‘somewhat’ positive, while only 15% said ‘somewhat’ (10%) or ‘very’ (5%) negative. One third of respondents (35%) said they have a ‘neutral’ view of Shared Source.

    {How deftly they dance around the fact that, after the pitch, at most 2% are ‘very positive’.}

    • Although not overwhelmingly positive, the majority of US (55%), Brazil (53%), and French (52%) respondents rated Shared Source at least ‘somewhat’ positive. The Japanese were the least convinced with only 30% of respondents rating Shared Source “somewhat” or “very” positive.

    • Non-IT BDM’s reacted most positively to the description of Shared Source, with 57% rating Shared Source as at least ‘somewhat’ positive. This was followed by IT BDM’s (50%), IT Pros (44%), Developers (43%), and Issue Elites (41%).

  • . After being read a series of possible Shared Source benefit statements, large majorities in every country save for France (41%) said that the Shared Source initiative offered at least the same benefits as OSS. Focusing in on the audiences, large majorities of every audience save for Issue Elites (40%) said that the Shared Source initiative offered at least the same benefits as OSS. Support for Shared Source was strongest in the U.S. (73%) and with IT Pros (71%).

    {Cleverly spun, but their own figures elsewhere show they couldn't get the post-pitch approval rates for shared source up to the pre-pitch approval rates for open source. And that's in a situation where Microsoft was able to manipulate the presentation of shared source, including trying a series of possible benefit statements. They don't give country-by-country figures for those “large” majorities, either. I'll bet they just kept promising more and more apple pie and motherhood until they could scrape up the figures management wanted to see.}

Open Source & Linux Messaging

  • Direct attacks of OSS and Linux are NOT highly effective. Messaging that discusses possible Linux patent violations, pings the OSS development process for lacking accountability, raises the specter of possible security flaws, and the like are only marginally effective in driving unfavorable opinions around OSS and Linux, and in some cases backfire. On the other hand ‘positive’ OSS and Linux messaging, i.e. access to the source code, the price, lower TCO, the ability to freely make copies, and the like drive very favorable opinions around OSS and Linux, both across geographies and audiences.

    • “Linux patent violations/risk of being sued” struck a chord with US and Swedish respondents. Seventy-four percent (74%) of Americans and 82% of Swedes stated that the risk of being sued over Linux patent violations made them feel less favorable towards Linux. This was the only message that had a strong impact with any audience.

    • Some criticisms of OSS backfire: Ratings for messages that were meant to be negative actually had a positive response among the respondents. For example, when read what was supposed to be a negative OSS message about OSS and proprietary software having a similar TCO, nearly half (49%) of all respondents said that having heard this message they were now MORE FAVORABLE towards OSS.

    • The most effective OSS positives focus on TCO and the ability to compete with the United States. The top rated messages for OSS among all audiences were that OSS was ‘Cheaper & allowed free copies’ (84%), followed by ‘Avoiding payment of royalties to US companies’ (81%), and ‘the opportunity to build local the local tech industry to compete with the US’ (76%).

Shared Source Messaging

  • The most effective Shared Source messages 1) offer a transparent benefit and 2) are audience specific.

    • IT BDMs and developers are encouraged by messages that indicate that Shared Source will make it easier to build applications based on APIs and that Shared Source will build the developer community.

    • Customers (non-IT IT professionals and BDMS) respond best to improvements in the feedback process, and being able to perform security checks.

    • Issue elites (with the exception of Japan) respond extremely positively to increasing education access.

  • Messages that rely on an abstract discussion of intellectual property rights are not effective.

  • {No surprise. The FSF pushed that line for 15 years and it didn't do us any damn good either.}

    • The discussion of IP rights needs to be tied to concrete actions.

    • {Uh oh. I have to wonder if ‘concrete actions’ is code for ‘massive #@%!$ing lawsuits’.}

    • Note that as with the International Government Elite Survey (IGES) project, here respondents do not see the connection between Intellectual Property Rights (IPR) and a strong technology industry.

      {Gee — I don't suppose this could be because they've figured out that the Microsoft/RIAA/MPAA version of "strong IPR" actually strangles technology industries for the sole benefit of predatory monopolists?}

  • Shared Source messages are effective.

    • After being read this series of possible Shared Source benefit statements, 60% of all respondents said that the Shared Source initiative offered at least the same benefits as OSS.

    • Focusing in on those respondents who said they are ‘mostly favorable’ toward OSS, 60% felt that Shared Source offers benefits that are equal to (40%) or better than (20%) OSS.

    • {This means that we need to be louder about what a legal death-trap ‘shared source’ is — how the NDAs and other IPR restrictions Microsoft has around it could open their partners and customers to a severe risk of lawsuit should they ever want to do anything that Microsoft interprets as a competitive threat.}

Summary

  • Overall, the greatest challenges we face are with the International audience — especially the French, Germans, and Japanese.

    • The French are looking for an alternative to Microsoft, have high familiarity and favorability of OSS and Linux, and a strong belief that Linux has a lower TCO than proprietary software. This geography, while not yet ready to broadly deploy Linux with their businesses, is very interested in OSS and its potential. The vast majority of this audience had not heard anything about Shared Source, but was more positive than negative towards the idea. They do not feel, however, that Shared Source will provide better benefits than OSS.

    • The Germans are not as familiar with OSS and Linux. However those that are aware have very high favorability of both OSS and Linux, and are very interested in broadly deploying Linux. In addition, a large majority believe that Linux offers a lower TCO. This audience had heard little about Shared Source, and was mostly neutral to the idea. However, after hearing about Shared Source, the majority felt that it could provide ‘about the same’ or ‘better’ benefits as OSS.

    • The Japanese are very familiar and favorably predisposed towards OSS and Linux. This geography is interested in broadly deploying Linux and does believe that it offers a lower TCO than proprietary software. While many Japanese respondents have heard something about Shared Source, this audience was mostly neutral on their feelings towards shared source and most felt it would provide ‘about the same’ or ‘worse’ benefits as OSS.

    {Read carefully, this suggests that they were not able to persuade a majority that shared source benefits would exceed open-source benefits.}

Additional Information (Survey Results/Individual Country and Executive Decks):

Research Contact: Kathryn Almendarez Marsman, Research Manager, kathalm