An interesting and unexpected subtext of a recent thread has developed in the comments area over how we determine what is and isn’t spam here at Hmm. A reader, Randy, thinks we were calling the company in question a spammer, when in fact we were calling what the company’s representative did in the comments area spam.

There is a huge difference there and one worthy of making another post trying to explain the difference.

But first, perhaps for some readers our comment policy might be unclear and I decided to revisit it and explain in more detail how we handle comments at this blog. I’ve written about this subject before in saying that we intentionally do not use rel=’nofollow’ on the comments and I want the search engines to follow commenter’s links.

However, I still do not feel trackbacks and comments are rights, they are privileges and we (site owners) are the gatekeepers who decide which ones come through and which ones don’t. Here is something not every reader may realize: every blog post made at this blog received dozens (and some receive hundreds) of comments. If you look at the actual comments displayed numbers though, you won’t see these numbers, because the vast majority of these comments are comment spam bots, not real people.

So it’s like sifting through your email trying to separate the wheat from the chaff.

The search engines send people here looking for relevant material, not chaff (spam), so we need to filter this out or the content becomes polluted. Search engines will penalize sites that link to spam and/or are filled with spam. At least that is the common belief, anyway. Since a significant percentage of this website’s traffic comes from search engines, we need to have some rules to monitor that this activity doesn’t get out of hand, so we can continue to enjoy a new flow of readers coming from the search engines.

Yes, we could put rel=nofollow on all comment and trackback links. That is the SE suggestion, but I don’t like using that blanket theory because I feel that good comments deserve to have the search engine follow their link. Good, relevant comments add to the discussion and might help the person that comes to this blog looking for information to ultimately visit the commenter’s site and get exactly what s/he was looking for. It’s sort of like that classic give the customer what he wants scene in Miracle on 34th Street where Macys sends customers to Gimbals. That’s a win-win-win situation. Win for us because the user came here from the SE and found a link to the information s/he was looking for and a win for the commenter because s/he gained a new site visitor and win for the SE because they were relevant for the search. The system worked!

What type comments are removed?
Generally speaking the following types of comments are removed when left in the Hmm comments area:

- a comment that only contains a link, unless the link is to a blog entry that expands specifically on the thread and the article/blog entry links back to Hmm. This is probably best done with a trackback.
- comments that say: hey check this out, or some variation, and include a link. Yes, there are exceptions where these posts can be approved as noted above. When they are from somebody who represents a competing company, such as the case that happened here, and then received the response here.
- one way trackbacks. Trackbacks that come from third party sites which write about the topic generally but do not link back to Hmm. In very rare instances when we feel the content is outstanding we will approve these trackbacks, but otherwise these are deleted.
- comments containing known spam keywords. There are a half-dozen or so keywords which are blacklisted from the comments area. They are keywords very common to comment spam and extremely uncommon to legitimate comments. Yes, we do realize some false positives may result here. If you left a legitimate comment and it didn’t show up as being approved, then please email that comment to tdavid at gmail dot com and we’ll add it for you manually. We aren’t trying to stop legitimate, relevant comments. Yes, even those that strongly oppose or disagree with something published at Hmm.
- comments which bring nothing of value to the discussion. Yeah, this is very subjective. If you want to ANONymously tell me how much of an asshat you think I am or how stupid this blog is then use your own blog and then trackback in. We’ll be much more likely to let that type of commentary passthru then a snappy ANON one liner. Still, there are exceptions and we do let some of these type comments through. Particularly if they made us laugh. We’re suckers for good humor and biting satire.

Bottom line is that we remove what we feel doesn’t add to the conversation in some meaningful way. This doesn’t stop at spam but extends beyond spam. Sometimes spam does add to the discussion in a meaningful way. It can display which people from which companies have class and show good netiquette. This can be valuable information in and of itself.

I will be the first one to admit that this is not an exact science. Is it fair? I think so, yes. Is it complicated? On its face, yeah, it seems like there are a lot of rules, so why bother, but the reality is if a commenter always does the following:

Submit comments or trackbacks (with attribute link) on topic adding useful commentary to the discussion — whether you passionately disagree or agree — and don’t use the half-dozen or so blacklisted words (if you run your own blog that has comments enabled then you can guess what these words probably are rather easily) your comment will always be approved.

Back to the issue Randy and I have been discussing in the comments area. That particular issue which I’m intentionally avoiding using the company or rep’s name (that’s been done enough) was originally identified not as spam but as a “risky move.” I was careful not to call that SPAM in the beginning because I knew it was a very subjective call. Towards the end of the review I wrote: “And I sure as heck hope your message wasn’t a spam being run out to multiple blogs.”

But the more I thought about that, after writing it, I changed my mind and believed it was, in fact, spam. Not the company spamming, but the individual spamming. I probably should have never approved the comment and sent the individual a private note not to leave those types of comments on this blog.

I still would have done the review though. And the content of my review would not have changed one iota. I think that service needs some serious work in the reliability angle. But did I ever call the company spammers? No. Did I call the actions of one of their representatives in the Hmm comment section a spammer? Yes, I did, in my follow-up post here.

Who cares what we think is or is not spam? Well, maybe only the commenter, Randy, I and the people interested in the discussion of what is considered spam and what isn’t in the comments section of a blog.

Moral of the story here is if you start your own blog and decide to accept comments you will eventually have a few important decisions to make:

Will you …

1) accept every comment, including spam
*2) accept every comment, except for spam — but how will you determine what is and isn’t spam?
3) accept every comment that is on topic, including spam
*4) accept every comment that is on topic, except for spam — but how will you determing what is and isn’t spam?
5) accept every comment that is on topic, including personal attacks, including spam
6) accept every comment that is on topic, but does not include personal attacks, including spam
*7) accept every comment that is on topic, but does not include personal attacks, and does not include spam

Yikes! See the problem? There are all these if, ands, and buts in the equation. The simplest and safest solution if you don’t want your site to be earmarked as spam by the search engines is to choose something like #2, #4 or #7 above (denoted by asterisk).

Personal attacks and trollish behavior are another angle that we can get into in the comments area, but generally speaking my attitude on that is sticks and stones …

• First comment thread at Hmm reaches 100+ comments and gets derailed [site news]
• Trackbacks now inline, mixed with comments
• Comment moderation dilemma
• Comments and Trackback Policy
• Blog refer spamming
• How to identify and remove Wordpress comment spam MySQL queries