After Matt Cutts asked for feedback about search quality, webspam, webmaster tools and what not on his blog, he has been totally swamped with suggestions. Some really good, and some that would really kill Google as a search engine.
I really had a blast reading John Andrews´. It sounds as if he wants a totally different search engine. More something of a seaarchable catalogue of established large industry sites:
1. Add a filter so any site with less than 100 or so links from the same industry never shows up in the SERPs. Let’s face it, less than 100 quality on-topic backlinks demonstrates very little effort to reward with a top ranking.
This suggestions would in fact break everything that Google ever was about. Google lists pages, not sites, and how many sites do you know with more than 100 backlinks from the same industry? Imagine how long it would take for an online store or manufacturer to introduce a new product and make it visible on Google…
2. Add a time-delay filter so new websites don’t get a shot at the first few (10?) pages of serps for something like 2 years. That should eliminate alot of the pressure for rankings. Most of the new website owners will forget about high-pressure SEO-style tactics if they know for sure they have no shot of ranking for that long. Those that are still good after the delay deserve to rank well.
Some people say that this is what Google already does. It is refered to as the Google sandbox. I believe it’s there, but I think it has a lot to do with incoming links and update frequency. A new site that keeps adding pages, and that aqcuires new backlinks won’t end up in the sandbox. Most others do…
3. Only count visible, on-page text for relevance. No tags at all, no scripts at all, no code, no images nor their tags, just good old text. Start from there and impose your algos. SERPs should be much better even as design/styles vary and technology “advances”.
Now even if this is a suggestion I can agree with, it would effectively kill off most of the design oriented sites, and probably make the web an ugly place to visit.
4. Stop valuing backlinks by PR. The cascading relevance thing is dead, and rewards spam networks and scammers. Why not just kill it? Eliminate all that nepotistic link love and watch the SERPs improve.
This is truely cutting at the core of Google’s algoritm. However, it seems that this is another thing that Google already does. The new buzzword is trustrank, not pagerank.
5. Require the Google toolbar of everyone who wants to use Google.com. Why not? It helps you keep Google.com clean, so it can be required. Then you can track all activity, and really know relevance. (Granted this won’t work until you own the entire web, but that’s just a little while from now so better get crackin on the toolbar insertion, no?)
Now that would get the tinfoil guys out there into a fit!
6. Require AdSense on all web pages in Google. Again, why not? The Internet hasn’t been non-commercial since the mid 90’s. Offer a profit share as an option. Google’s revenue will go up, and you will eliminate much of the one-sided economic appeal of AdSense spamming, because every bit of content in Google.com will already have ads on it. No need to scrape sites to put ads on at another domain. Right?
Ha-ha! Another blow at Google, who claims AdSense and natural SERPs have no connection what so ever.
7. Change the Google indexing algorithms randomly every 5 months or so, on schedule. By my read, each big change in Google leads to about 4 1/2 months of whining in the spammer forums. Also, in my experience most users don’t know a good SERP from a bad one… they just click the first few results no matter what. Those two facts support a plan to just whack the index, and then whackit again as soon as the SEO spammers stop whining and start scheming again. Whack, whine, scheme, and Whack! again. Spam problem solved.
I actually thought this was happening already…
Still. I think Google would actually like to do most of the above, but it is so totally politically impossible today.