Recent data from an Irish metrics firm shows that the free upgrade to Windows 10 initially kick-started usage of the new operating system, but that within a month of its July 29 release, gains fell below the threshold set by Windows 7 six years ago.
Although the numbers reflect only the first 40 days of Windows 10's wider availability -- and the early burst of increases fueled a record for Microsoft -- they illustrate a shorter-than-expected impact of the free upgrade.
Data from StatCounter, which measures usage share, a percentage of the total desktop-and-laptop-driven online activity by a specific operating system, was used by Computerworld to analyze Windows 10's gains. Week-over-week increases in usage share for both Windows 7 and Windows 10 during their respective first 40 days were calculated, and then to eliminate daily variances -- usage of a new OS typically spikes on weekends because consumers compose the bulk of early adopters -- the results were charted using seven-day rolling averages.
The results were surprising.
Although Windows 10's average week-over-week increases started out extremely strong -- in the neighborhood of 2.4 percentage points -- they were halved by Day 21. Eight days later, Windows 10's increase was below Windows 7's at the same post-release point. Since then, Windows 10's usage share gains continued to slip, reaching four-tenths of a percentage point on Day 40.
Comparisons to Windows 7 were also illuminating.
That operating system was a traditional release for Microsoft in that it was a paid upgrade for those who wanted to install the new software on an existing machine. Microsoft sold discounted upgrades for Windows 7 for less than two weeks prior to its October 2009 launch, pricing those upgrades at $50 (Home) and $100 (Professional). The bulk of Windows 7 gains, however, came not from the in-place upgrades, but from users purchasing new systems.
In light of the long-term shipment slump of PCs -- three and a half years and counting, with no immediate end in sight -- Microsoft felt it necessary to provide Windows 10 free of charge in order to jump-start usage. It simply couldn't wait for the usual replace-with-new-PCs model to gradually push customers to the new OS.
Windows 10 has continued to increase its usage share, according to StatCounter: On Sunday, the OS broke the 8% mark for the first time. And its lead over Windows 7, thanks to the early surge, remains secure. As of Day 40, Windows 10's usage share was 1.2 percentage points higher than Windows 7's, representing a 44% advantage.
But the fact that Windows 10's average week-over-week increases have been steadily shrinking, and and more importantly, that they've dipped below those of Windows 7's for the last 6 days -- and in 8 of the last 12 -- signals that the biggest results from the free upgrade are behind Microsoft. Going forward, increases will likely be more in line with historical patterns.
The comparison between Windows 10 and Windows 7 is not apples-to-apples, maybe not even apples-to-oranges, since there are dissimilar variables, ranging from the time of year. Windows 7 at 40 days was into early December, PC buying prime time, while August is at the tail end of the second-best stretch, back-to-school, with few Windows 10-equipped devices yet on shelves. And there are vastly different personal computer dynamics in 2015 versus 2009.
Sans a free upgrade, however, Windows 10's increases would have likely been much, much smaller, far below Windows 7's, putting Microsoft in an impossible spot. But that counterfactual is impossible to prove.
One thing does seem clear: If Windows 10's usage share -- and data derived from other third-party sources -- doesn't improve at a faster pace, Microsoft could run through the 12 months of its free upgrade without enough to justify the effort, or make its professed goal of 1 billion Windows 10 devices by mid-2018.
Computerworld will check in on the progress of Windows 10's usage share gains at intermittent intervals.
This story, "Free upgrade gave Windows 10 a one-month boost" was originally published by Computerworld.