There are four basic methods used to categorize and filter Internet content...
- Manual - humans perform research, find new content, then review and categorize it. Results are added to filter database.
- Automated - specialized computer programs find new content, then categorize it, primarily by recognition of keywords but also by using increasingly more complex and sophisticated algorithms. Results are added to the product's content filtering database.
- Automated-plus-Manual - specialized computer programs find new content and provide a preliminary category assignment. Humans review the preliminary assignments and make corrections before results are added the product's content filtering database.
- Dynamic - specialized computer programs dynamically examine Internet traffic. Content is evaluated in real-time, primarily by recognition of keywords but also by using increasingly more complex and sophisticated algorithms. No underlying content filtering database is required.
Let's take a look at some of the positives and negatives of each method...
Manual: Positives
- Best recognition of context, leading to best category decisions.
Manual: Negatives
- Category decisions can be subjective.
- Slow and costly.
- Economically impossible to keep up with rapidly expanding universe of Internet content using manual method alone.
There are no longer any quality solutions that rely solely on the Manual method. All modern solutions rely (at least to some degree) on Automated or Dynamic content recognition.
Automated: Positives
- Category decisions are objective.
- Fast and relatively comprehensive.
Automated: Negatives
- Even the best programs are still not perfect at understanding context.
- Tendency to both over-block and under-block.
There are a few solutions available that rely solely on the Automatic method. Unfortunately, the current state-of-the-art is, in our opinion, still not good enough to eliminate manual (human) review. It is possible that a quality Automatic-only solution will become available in the future, and we constantly evaluate new product offerings. Currently, there are no products based solely on this method that we can recommend.
Automated-plus-Manual: Positives
- Best recognition of context, leading to best content category decisions.
- Fast and relatively comprehensive.
Automated-plus-Manual: Negatives
- Category decisions can be subjective.
This is the preferred method used by the leading providers. It combines the strengths of both the Automated and Manual methods. Automated computer programs handle the sheer volume of new content, preprocess the results for the humans, who then provide the Manual final review to ensure correct categorization.
Dynamic: Positives
- Category decisions are objective.
- Comprehensive, since all content is reviewed in real-time.
- Continual updates to a content filtering database are not required.
Dynamic: Negatives
- Even the best programs are still not perfect at understanding context.
- Non-textual content is more difficult to categorize.
- Acceptable performance for high bandwidth applications is difficult to achieve.
There are several solutions available that rely solely on the Dynamic method. Unfortunately, in our opinion, none of them are yet accurate enough to eliminate manual (human) review. Currently, there are no products based solely on this method that we can recommend.