smithaa02 - 10:20 pm on Mar 4, 2011 (gmt 0)

Really hate how google did their human sampling... Willingness to hand over credit cards and receive prescriptions relate to ecommerce...but ecommerice is a small percentage of the web! Most websites don't directly sell things.

So without realizing it google with their stupid tests has rewarded large websites that sell stuff. When I use google, I 99% of the time don't use it to buy something but to learn something new. The big 'status-quo' name-brand companies and our 'trusting government sites' actually don't contain that much new information...it's actually the smaller more independent sites (like this one) that tend to contain information one wants. This is because independently produced content is of higher value.

So the result is the likes of Wal-mart and Amazon are huge gainers even though we already know all about them and they don't really have that much new to share. Recently have done a lot of private research and I'm getting a ridiculous number of non-relevant shopping sites that are muddying up the results that I didn't see before. "Appeal to Authority" is a common debating fallacy but apparently at google it makes for good SERP's :( Matt doesn't seem to realize that just because a site isn't a government site doesn't mean it can't have better information about "pediatric multiple sclerosis".

As for being so proud their results matched "chrome site blocker" this is absurd... The web is a huge spectrum of quality in which these big abusers are a small percentage. To emphasis this 'fix' across the spectrum means, yes they will knock out the bottom say 1% of sites but will drag down a significant percentage of sites on the 2-40% end of the scale as well. It's like google asking "chrome site blocker" if chopping off the foot will catch a wart on the pinky toe..yes it will but it misses the point. Now what do the non-farm collateral damage sites and actual farm sites have in common? They apparently don't have that big corporate/government structure backing...could be ridiculous domain ages / page count / site structure or a number of other ridiculous non-content variables that really don't pertain to quality.

Sorry for rant...just can't believe how clueless some of these google engineers are when this should be stats 101 :(