Future of Search Engine Results
What is the end result for search engine rankings? My theory! Currently the search engine algorithms are based heavily on link based methods. These methods are continuously being manipulated, and regardless how the algorithm evolves or updates, the manipulation will continue. What are the solutions for the search engines? Well, one that comes to mind is the use of directories. I know what you're probably thinking already, been there, done that, webs gone forward, however; let's put a new spin on things shall we.
The current system is flawed, regardless where it turns. Search engines continue spending countless hours on human resources, technology and so forth. They introduce multiple algorithms, all of which can be manipulated, regardless how often they change them. What is less susceptible to manipulation when controlled by the search engines themselves? A directory!
Now, don't get me wrong here with technology, as it is a vital key to the success of all things web, and it definitely has its role in this system, as I will discuss a little further on. If each search engine had its own unique directory, then a system of random results could be determined from relevant categories when searched by the user. The directory would have to be owned, built and maintained by paid employees of the search engine, otherwise it is immediately open to corruption, as like DMOZ is now finding out. Now we all know what a directory is and isn't, so we don't need to start at the basics.
Currently when SE's return results, they are based on a scale of probability assessed by multiple algorithms, and computed on a weighting of those factors to determine relevance. It is these factors that are continuously being manipulated by the world of webmasters, to prevail themselves as the winner. Now taking this into account, lets take the current result and user habits into account.
Conducting a search for "real estate" on Google, returned over 299 million results. Who is ever going to look through that many results? Nobody, that's who. Out of that 299 million, do you really think they are all related to real estate? I doubt that very much. Any page that lists the word "real estate" would be in those results basically, thus rendering an enormous proportion of those results useless to the user. Continuing to use this as an example, with geo-targeting technology built in, if you typed "real estate" into google.com, then you would expect only real estate sites that list across the entire US to be returned, same same if you typed real estate into google.co.uk, you would only expect sites that list real estate across the entire UK.
With localized search already implemented, if a user types in "real estate new york" or "real estate NY", etc etc, then they should only be returned results that list properties for sale in New York only, not the entire USA. A directory categorization, in conjunction with geo-targeting is already sounding pretty good. If the particular New York real estate category listed 5000 sites, then that is much better than the current 88 million results returned.
Now, 5000 sites is still out way out of the average users scope to sift results, though atleast the results are narrowed down to the actual listings, and not every page on earth that currently mentions the term "real estate new york" or "new york real estate". Going another step further, in conjunction with semantic algorithm technology, each of these 5000 results could be further refined to show only those pages that are semantically correct, and ascertained as direct New York real estate listing sites / pages. Lets say we knock out another 1000 results with the algorithm refining results for us, leaving us only 4000 returned pages for the term. Personalized search could then become a large factor from the monitoring of user surfing habits and refinements, knocking out sites that contain less useful information to their query.
Saying that, personalized surfing habits are left a little wide open for the use of software to manipulate surfing habits throughout key sites, thus making it a little doubtful. So the search engines need a little extra something to refine those searches for the user. This could simply be achieved by knocking out non-local results. For example, if a directory category is "real estate new york", then under that category would be all the specific regional areas. If "x" amount of results are returned, then the SE needs to ask the user to refine their results. So, give the user the choice to select a regional suburb / locality within the larger scope.
By doing this, a user is then forced to input more data to find the relevant information, all the while, 100% within the search engines control, and still not open to manipulation of any kind. So lets now say that the user is actually looking to buy a property in Queens, so the search engine has forced the user to input something more along the lines off "real estate new york queens", which now narrows the results to only 600 agents, for examples sake.
As you can see, through directory categorization, search engines could actually return much better results, without any manipulation from a webmaster, as current algorithms already determine most of these factors, its just they are flawed by the use of links for weighting purposes.
The problems fixed via this method far outweigh the current trends and movement of the algorithms. A specific analysis program that assesses each sites credibility as such, for inclusion with a specific category, would help editors to most of their work for them. If such software could analyse one site per minute, at a small fee of say $5 per site, this means 60 sites per hour, 1,440 sites per day, 525,600 sites per year, for one employee. I'm not saying that search engine editors need to view each site, as that would be sheer stupidity, and not remotely possible to achieve, though algorithms could review a site on a points system, for its credibility to a specific category. If it fails, you resubmit, after narrowing your category selection a little further, or broadening your site.
So what would happen then? Well, every site owner would be building site after site in order to get more listings within each category. I envisage you could take out the stupid, those that link to the main site, or those that network their domains for usability. That would take out some. Lets face it though, this is already happening, in that businesses are buying hundreds, even thousands of domain and hosting them with slightly different content, or breaking their content up to specific groups, per domain. Not sure how to solve that one exactly, but I'm sure it could be narrowed quite dramatically.
This would make linking obsolete, so all these pages and sites that exist for the sole purpose of linking, would all be gone. Well, those that are for SE ranking purposes only. Let's face it, if the tables turned and the web became all about usability, instead of manipulation, then people wouldn't have half the junk on their sites they do now. This would cut the overall clutter of the SE's down quite drastically IMHO, even taking into account practices that are already occurring with multiple domains.
Cost is a limiting factor for most, and domains could still be easily identified through DNS searches, registration names, business tax file numbers (if applicable) and so forth, to really cut the extra domain down to a more holistic figure. These directory structures and categorization would definitely take the manipulation out of the hands of the webmaster, and place the rankings (for the most part) back in the SE's hands.
Its a theory, and like all theories, mean very little more often than not.
