Does Microsoft's "Forgetting" the EU's Browser Ballot Matter Anymore?

The world once debated whether Internet Explorer's dominance in Web browsers was fair in the wake of Microsoft's conduct against Netscape. As part of its 2009 settlement with the European Union, Microsoft offered to give new Windows users in Europe a choice of default browsers, rather than just leave them with Internet Explorer and make them manually install any alternatives. At the time, one objection to this plan came from one of the very organizations whose objections to Microsoft's conduct led to the settlement in the first place. The European Committee for Interoperable Systems (ECIS) alleged that the choice itself would be perceived by users as an annoyance at best, and at worst, a threat.

The ECIS may have been right.

Tuesday's revelation that Windows 7 Service Pack 1 in Europe omitted the browser ballot for 17 months - without anyone raising a fuss - suggests that users may have been better off without it anyway. As Brian Proffitt reported yesterday for RWW, Microsoft indicated to antitrust representatives of the European Commission (the upper house of the EU parliament) that the ballot was present in Win7 SP1, when in fact it was not.

European Commission Vice President Joaquin Almunia Tuesday morning continued to take credit for the success of the browser ballot initiative as though it existed. "To avoid the tying of Internet Explorer... in December 2009, the Commission made legally binding on Microsoft a commitment to make available a 'choice screen' enabling users of Windows in the EU to easily choose their preferred browser," Almunia told reporters in Brussels. "This decision, therefore, put Microsoft under an obligation to provide the choice screen to European Windows users until 2014. This has proved very effective when implemented: It gave consumers a real choice to use the product that most suited their needs."

No Effect From Choice Screen?

But evidence from one of the world's leading analytics firms suggests Google Chrome's rise to Europe's number one renderer of Web pages was not impacted by the ballot screen's absence. StatCounter's continuing measurement of which browsers render Europe's HTML pages, conducted since before the EC okayed Microsoft's settlement terms, suggests that the decline in IE's usage share for Web page rendering continued at an even pace even after the choice screen was neglected.


Source: StatCounter Global Stats - Browser Market Share

It's important to take into account how StatCounter measures usage share (which is often inaccurately confused with "market share"). StatCounter does not count the number of installed browsers. Rather, it tracks browsers that identify themselves as clients rendering requested Web pages. (StatCounter recently made an adjustment in its scoring process, to take account of the fact that Chrome pre-renders pages that are linked to the visible page - a fact which had previously given Chrome an undeserved edge.) So among the top five browser brands in Europe for June 2012, Google Chrome (all versions) now stands tied with Mozilla Firefox (all versions), each rendering 30% of the continent's pages. IE (all versions) is responsible for 28.19% of Europe's page rendering.

A Complex Failure

To say that the Windows 7 SP1 installation process failed to include the browser ballot is slightly misleading. In fact, the installer never did contain the browser ballot, but instead pointed to the site where the ballot is hosted by Microsoft in concert with the EC. Rather than be led by a Microsoft-managed process, was originally intended to be an independent service run on Microsoft's dollar.

The site displays links to the top five browsers, as well as to seven more (when you scroll right), along with advertisements for those browsers, in a random order. The decision to randomize was made after Mozilla suggested that the original order (alphabetical by manufacturer) would have been deferential to Apple, which releases Safari. Even the formula used to randomize that order came under suspicion when IBM software engineer Rob Weir, in private testing, discovered that the algorithm Microsoft chose for shuffling browsers' positions on the ballot was unfair - ironically - to IE.

Separately, there were whispers of a conspiracy by Microsoft to monopolize the browser screen by effectively giving some of the seven "secondary" browsers (the ones you see when you scroll to the right) the IE rendering engine, enabling them to report themselves as IE to services like StatCounter. The belief that the browser choice screen may actually have been a subliminal ad for IE triggered a grassroots movement by some in IT to disable the patch that delivered the link, including through the use of registry hacks.

Waste of Energy?

All of this is trumped by the realization today that the ballot never made much of a difference to users in the first place. Usually Windows service packs include "roll ups" of patches that were released heretofore. Win7 SP1 for Europe may have omitted KB976002, the patch that contains the link, which was distributed to all Windows users.

Almunia stated that the EC is opening formal proceedings against Microsoft for essentially failing to include the link as part of the rollup. But that failure actually might not have precluded all Windows 7 SP1 users from seeing the link. Conceivably, they could have downloaded KB976002 as part of their regular updates, for the reason that it was not included in SP1. Thus, European users may have actually been shown the ballot anyway, even though the SP1 process omitted it, by virtue of an online check that may have taken place once the local installation process was concluded.

The amazing thing is the revelation that apparently nobody actually knows whether users got the ballot or didn't. Although Microsoft is ultimately responsible for the truth of its own statements to the European Commission, conceivably the EC could have easily checked on the relative popularity of the site using its own means, had it bothered to do so. Some factors regarding the site's popularity (which is arguably quite low) are, and have been, in the public record.

Windows 7 Ascendant

During the browser ballot's absence, Windows 7 was surpassing Windows XP to become the operating system behind the browsers rendering more than half of Europe's Web pages, according to StatCounter's estimate for last month.


Source: StatCounter Global Stats - Operating System Market Share

The statistics seem to indicate that the precipitous rise in Chrome's usage share is attributable not to the kind of filtration the browser ballot was intended to achieve, but rather out of users' free will. It will be interesting to see whether investigators take note of whether users now choose Chrome because they like it better or, as one Mozilla engineer suggested in 2010, because users confuse Chrome for Google itself.

Put another way, if Microsoft owes the public restitution for IE's preferential treatment in 2009, what does Google owe the same public in 2012? And suppose Google were made to give the public a choice... Would anyone notice?