: What value do comments and ratings have anymore? I am developing a hit-and-run blog style posting mechanism for my already otherwise mind-bending HTML short-hand CMS schema that runs several of

@Sherry384

Regarding Comment Sections:


Search terms used: comments section seo


There are several considerations that need to be made before comment sections and ratings are to be implemented.


General:



Time it takes to moderate the comments
Moderation must happen quickly or comments section will suffer
Whether the comments reside on the site itself
Whether schema.org mark-up is used - no evidence
Frequent changes to the page helps keep the page fresh - no evidence
Whether lazy-loading is used - may not be read
Whether pagination is used versus all on one page
Can help to support otherwise thinner content, lazy content creation
no evidence
Added linguistic semantic value, so-called LSI keyword variation - no
evidence
Broken links and management
Me too, great post, or other useless posts must be removed
Grammatical errors block proper semantic analysis and can lower page
quality
Lack of comments appears negatively
High authority sites may not benefit from comments - no evidence



My Comments:


Evidence is largely anecdotal. No real evidence or metrics are offered.

LSI keyword variation is primarily a theory of synonym use and that the use of synonyms are valuable. This is backwards logic considering the evidence of synonyms supports Googles use of synonyms in search and not how synonyms add value to a site.

LSI = Latent Semantic Indexing

It is not necessary for content to include synonyms since LSI will happen within the index anyway. Google recognizes the importance of recognizing search terms not used but still apply and will give semantic weight to any term that applies and exists within the synonym ontology. In this, Google equates synonyms automatically and using synonyms is not generally necessary and is a trigger in spam analysis.

My evidence supports that synonyms are valuable, however, not the way you think. It is wrong to use synonyms to add signals that Google does not need such as car, automobile, vehicle. However, it is correct that synonyms are used to directly support semantic weighting of topical terms, nouns, pronouns, etc., that are not generally supported within the ontology that defines synonyms, but rather are supported within topical ontologies. It is also not recommended that a content writer use synonyms for the sake of using synonyms and that a semantic analysis tool be used to weigh the effectiveness and appropriateness of any topical synonym. This works best on scientific or technical topics. For the record, there are cases of semantic term classes that synonyms work well for to support characteristics of a written work that while is not necessary from the standpoint of synonyms within an ontology, do support other aspects that positively effect search for other reasons. This is where the notion of synonyms by SEOs is wrong-headed and understanding the effects of synonyms far more technical and subtle than ever imagined.

Comment sections can stand in the way of appropriately weighing your content using semantic analysis and linguistic disciplines.


Matt Cutts in 2013



Whether the usernames used for comments are real user names
Whether the real users are well known experts or quality posters
Whether avatars are allowed
Whether link backs are to authoritative sites
Whether the comments are from users who comment on many or several
sites often as a link back strategy
Whether the link backs are deceptive or become deceptive later
The quality of the outbound (external) links of the page overall
(ratio)
Whether comments is the primary outbound link scheme
The quality of the comments overall
Whether the comments are topically relevant
Whether the comments are using known spammy terms



My Comments:


The consensus on an unnamed and well known blackhat SEO site is that comments do not help and that all comments are spam. This may be true to a large extent and certainly the primary motivation of leaving a comment is promotion of another site. Another point made is that comments are not good for Adsense. How that manifests, I am not sure spefically, but it is easy to see how comments can negatively effect earnings or whether an ad is placed at all.

The first thing that is needed to understand is that comments are not content in the sense that Google can easily separate content from headers, footers, sidebars, and comment sections. This is done using the HTML DOM model of several pages using pattern recognition to determine templated content from non-templated content.

Much can be easily evidenced by doing a search for the overly used great post which may match a comment in a comment section. What surprises me is that these low quality searches match low quality comments. Here is an example that I found humorous. On this page, blog.kissmetrics.com/double-your-social-media-traffic/ the term great appears 31 times including 7 times within the article.

This prompted an experiment using the linked article. Please stick with me as I offer some evidence using Tropes.

A Semantics Example - Article versus Comments

On the outset, this is an article on a site where the writers and commenters are rather professional and intelligent. This is a very well moderated site. It is expected that this example is a more closer to a best-case scenario.

Article Text Style

Style rather enunciative
Setting: involving the narrator
Some notions of doubt has been detected


Article Reference Fields

communication 0030
time 0026
media 0018
computer_science 0008
business 0008
language 0006
social_group 0005
person 0004
money 0003
cognition 0003


Article Frequent Word Categories

* Verbs :
Stative 36.2% ( 138 )
Reflexive 25.5% ( 97 )
* Connectors :
Condition 14.6% ( 13 )
Disjunction 12.4% ( 11 )
Comparison 13.5% ( 12 )
Time 9.0% ( 8 )
* Modalities :
Time 29.2% ( 42 )
Manner 25.0% ( 36 )
Doubt 2.1% ( 3 )
* Adjectives :
Objective 69.3% ( 183 )
* Pronouns :
"I" 13.4% ( 16 )
"We" 13.4% ( 16 )
"You" 45.4% ( 54 )
"They" 7.6% ( 9 )


Comments Text Style

Style rather enunciative
Setting: involving the narrator
Setting: involving with "I"
Some notions of doubt has been detected


Comments Reference Fields

communication 0099
time 0099
media 0028
language 0023
computer_science 0015
work 0015
business 0013
space 0011
social_group 0010
device 0009
cognition 0008
location 0008
law 0006
feeling 0006
person 0003
characteristic 0003


Comments Frequent Word Categories

* Verbs :
Stative 37.0% ( 221 )
Reflexive 29.8% ( 178 )
Performative 1.0% ( 6 )
* Connectors :
Condition 10.8% ( 16 )
Disjunction 8.1% ( 12 )
Comparison 13.5% ( 20 )
Time 6.1% ( 9 )
* Modalities :
Time 26.8% ( 62 )
Manner 17.7% ( 41 )
Assertion 10.8% ( 25 )
Doubt 1.3% ( 3 )
* Adjectives :
Numeral 43.8% ( 283 )
* Pronouns :
"I" 38.2% ( 108 )
"We" 9.5% ( 27 )
"You" 21.9% ( 62 )


Here are some interesting sample points difference just for fun.

Comments Text Style

Setting: involving with "I"


This means that the comments lean toward the "person" with personal pronouns rather than being purely stative.

Reference Fields

The following increases Comment versus Content Reference Fields

communication 0099 versus 0030
time 0099 versus 0026
media 0028 versus 0018
language 0023 versus 0006
computer_science 0015 versus 0018
work 0015 versus 0000
business 0013 versus 0000


This list indicates stronger semantic values for topics communication, time, media, language, and computer science and lesser values for work and business. Please note, this is not a complete list, however, you can easily see that semantic value of the original article changes with comments both positive and negative. If the entire lists were to be analyzed, we could see a larger negative effect.

Space: Terms found; Moon as in Garrett Moon
Law: Terms found; traffic
Feeling: Terms found; pleasure, hope, friends, love, cheers


This illustrates how semantic topics can easily be swayed.

Frequent Word Categories

The following increases Comment versus Content Reference Fields

Reflexive 29.8% ( 178 ) versus 25.5% ( 97 ) the effect is negative
Disjunction 8.1% ( 12 ) versus 0% ( 0 ) the effect is negative
Comparison 13.5% ( 20 ) versus 0% ( 0 ) the effect is positive
Manner 17.7% ( 41 ) versus 25.0% ( 36 ) the effect is negative
Assertion 10.8% ( 25 ) versus 0% ( 0 ) the effect is negative
Doubt 1.3% ( 3 ) versus 2.1% ( 3 ) the effect is moot


While this is not a fully exhaustive analysis of the semantic difference between the original article and the comments, there is a lot more I can add but would make this post far too long. It is clear to see that comments can effect how a web page is found both positive and negative with the potential for the negative to outweigh the positive. And that is the point.

Back to Comments

Of the research I read years ago, a lot of work regarding comments pertained to usernames, links, and recognizing the quality of the and reputation of the poster, the comment, links, commenting habits, etc. Clearly Google is looking hard at comments when comments can easily be ignored. This indicates two things. One, Google found that comments can potentially offer value in site/page evaluation both positive and negative. And two, comments require a lot of scrutiny.

Regarding Article Ratings (stars):


Search terms used: article rating seo


No evidence found.


My Comments:


Evidence given in SEO articles seems to apply to product reviews or business reviews and not article reviews.

I did a few searches on Google for various SEO articles. Not one of them had reviews.

Clearly, the results indicate that using review mark-up for articles is not shown or not used by content creators. Without digging further, which is the case is not clear. However, the evidence is clear. Ratings for articles have no value.

Answering My Own Question:

Comments are a lot of work. Especially when you consider that I code my own CMS and 3rd party plug-ins are not available. Creating a plug-in (in-line forms process) is very possible, but may take more work than I wish to undertake at this point with questionable gain. As well, I am a busy fellow. I do not have the time, attention, or will to moderate comments. I may in the future.

Article reviews appear not to have any value. While mark-up is an important factor, schema.org does offer options without having to resort to article reviews. I may implement this feature in the future.

For now, I will not include these options in my schema. There is room for them in the future if the search winds change. While it is better to build schemas and code early for future options, for a hit-and-run application to post to my site quickly, the implication is that the work currently undertaken now be quick and effective.

Vote Up Vote Down

@Looi9037786

Value is already opinion-based term:) The value of comments you can benefit from is directly depending of the quality of your readers:


Are readers nice people - they write meaningful comments, you don't have the pain to struggle with trolls -> overall profit.
Once you have more pain/time loss on troll management, as benefit from meaningful comments - overall fail.


Google, with its overvalueing links, has perverted the comment idea - the most of web folks (my personal guess: 99,9999%) are commenting with the single goal of link dropping.

A kind of nearly clear answer to your question would be: yes, high quality user generated content has notable SEO benefit. But its value must go hand in hand with rational time investment to manage it.

It isn't only my personal point of view on the subject: many SEOs i met formulate it exactly so: yes, good comments bring SEO value, but the should always stay in relation with invested management time.

Edit:
assuming there are no trolls, comments would give some pretty concrete benefits:


additional unique content, coherent to the page's main topic and containing keywords from the same lexical core as the page's topic,
every comment adds a bit to the freshness factor, which is a ranking factor, specially in news-alike topics,
each comment gives a boost to the page's user metrics - more time on site, less bounce, deeper click through

Vote Up Vote Down