July 19, 2006

Want to rank sites by quality? Check spelling and grammar!

Hi Folks,

Google is doing everything in their capacity to weed out spam and return meaningful results for search queries. However it seems that they missed on an important aspect!

Concept

A good quality website will definitely have proper (if not accurate) spelling and grammar. Using advanced spelling and grammer checking routines, it is very much possible to weed out spam and provide higher rank to high quality websites.

Benefits

This approach has many advantages:

1. Quality will get preference over quantity

2. Sites which are ranked lower due to poor quality spelling and grammar have a chance of imporving ranks by correcting spelling and gramatical errors. This will initiate a rush to improve user experience and we will be able to see better quality sites all over.

3. Sites which use more “generic terms” and less “proper nouns” will get higher rank as they are simple to understand and are written with a generic audience in mind.

4. Spam sites which simply puts in pages generated from search results will get totally eliminated as they will have broken sentences.

5. Links farms, FFA, Generic directories can be identified and ignored unless specifically requested by the user.

Implementation

The implementation can be further enhanced by setting up a baseline and quality benchmarks, just the way Google did for link popularity (Page Rank).

Sites can be evaluated on a regular basis based on the following parameters:

  • Total number of words in the page
  • Spelling errors per 100 words of content in the page
  • Gramatical errors per 100 words of content in the page
  • Another parameter which can be useful is how the page validates. Is the page full of HTML errors? Is it XHTML validated? Is the CSS validated?
  • How often the page have been updated?
  • How often the site has been down?

This data can be collected over a period of time and can be be used to determine how the site has improved or declined in quality.

Can it be implemented?

Now the question is - “How difficult is the implementation?”

I have recently been watching some of the changes introduced by Google in order to refine their ranking algorithm:

a) Penalty being imposed for duplicate content
b) Reciprocal links getting lesser weightage than one-way links
c) Usage of Latent semantic analysis (LSA) to find relevant related results

If we consider the total processing power required for each of these refinements, we can safely assume that they have enough processing power to implement spelling and grammar check.

I will not be surprised if Google is already working on this. Therefore it is advised that try to improve the above mentioned parameters before the magical google update strikes you!
I hope to see “quality based ranking” in action soon.
Abhishek

Permalink • Print • 2 Comments