It drove me away.
Besides the horrible readability of the site now, my main method of news surfing (middle clicking on anything that looks interesting, and then going through each of the tabs one by one) no longer works on gawker sites.<p>I never realized how enslaved I was to gawker (kotaku and lifehacker to be precise) until I stopped reading them. The amount of free time I have now (to waste on HN) is astounding.
Part of the catastrophic drop may have been google referrals. I've noticed in the past with old sites if you do a dramatic upgrade on them, google recoils a bit and starts re-indexing. i've seen search referrals take up to six months to completely rebound. The big shocker here is if google hits their homepage, it's met with a big "fuck you". I knew they were doing "shabang" but didn't realize they had zero fallback for no js.
The SiteMeter graph is telling. But the Quantcast graphs, on which this article is hanging its hat, say nothing at all. The author appears to have misinterpreted the graphs without first looking up the concept of "overfitting".
Nice hack to get the old Lifehacker and Gizmodo back:<p><a href="http://ca.gizmodo.com/" rel="nofollow">http://ca.gizmodo.com/</a><p>or<p><a href="http://uk.gizmodo.com/" rel="nofollow">http://uk.gizmodo.com/</a><p>Works exactly the same for lifehacker. The new design, apart from it's disorienting scrolling behaviour, it's also very inefficient in terms of using screen real estate well. Impossible to quickly scan articles and determine which ones are worth reading. Fortunately you don't even need to see the ugly redesign.
The new layouts for both gizmodo & lifehacker seem quite buggy to me (Mac, latest Chrome). Most obviously, there is a grey line that tracks its way up the screen as I scroll past 'the fold'. And I agree with earlier posters: user test before release (qualitative), and split test as part of the release (quantitative).
Could this also be related to how data is reported to Quantcast? Many of what used to be page loads are now AJAX requests - and the page loads on Quantcast may have recovered after they added API calls on notify Quantcast.