Intel previously said that it expected its performance impact should not be significant for average computer users, and it thinks the data it is sharing today supports that expectation on these platforms.
The 6th generation Skylake-S platform shows a "slightly higher impact" approximately eight percent on the SYSMark2014SE benchmark.
Intel also measured performance on the same platform with Windows 7, a common configuration in the installed base, especially in office environments. The observed impact is approximately six percent on the SYSMark2014SE benchmark. Observed impact is even lower on systems with HDDs.
What Intel is not saying is that those figures are a lot higher than it expected to see for the newer chips. However the newer chips should hardly have been affected. On Skylake and beyond, Intel has refined the instructions used to disable branch speculation to be more specific to indirect branches, reducing the overall performance penalty of the Spectre "mitigation".
Older versions of Windows have a larger performance impact because Windows 7 and Windows 8 have more user-kernel transitions because of legacy design decisions, such as all font rendering taking place in the kernel.
Older chips which lack Skylake improvements will be gutted by the patches. We were expecting to see drops of 30 percent and yet Intel has failed to mention these older chips in its announcement. This figure is more important than the post-Skylake chips because there will be a lot more older Intel chips out there. Intel has promised to release figures for chips that are up to five years old which should cover the bulk of these older chips, but five years is no longer a long time in computing terms. There are a lot of PCs out there which are older than five years old.
Here are Intel's figures