Battling Comment Spam — The Big Guns

I was thinking about this comment spam series that I’ve written and I realized that it isn’t complete yet. There is at least one more post in this series.

You see, the series was written from my perspective. I’m not an A-List blogger and I don’t have to deal with the massive amounts of comments, email, etc. that they do. So, do the techniques that I’ve outlined scale to the rockstar level?

A-List Bloggers

I decided to take a look at what the big guns of the blogging and internet marketing niche do with their own blogs. I looked at some of the major sites to see what kind of things I could learn about comment spam from them. I also contacted several with a set of interview questions. I got 2 responses (thanks to the two Chris’s — Chris Garrett and Chris Guillibeau).  We’ll get to that in the 2nd half.

Update: I’ve received a delayed response from another of the big guns — thanks to Yaro Starak for taking some time from his busy schedule to help us out here.

Blogs Turning Comments Off

One thing that I observed was that most of the A-Listers still accept comments on their blogs. There are a few notable exceptions but the major players have comments in play — for most of their posts if not all.

Comments Closed

Another observation that I made was that some of them have comments closed after a certain time period. I’m not sure why they do this since their posts still rank highly in the search engines and I often get Twitter links back to those posts but I’m unable to comment on them. It closes the door to further discussion.

I suppose that older posts tend to get a higher spam to quality ratio. And the blogger has probably written 50 or so articles since then — time to move on. But it still seems like a bit of a wasted opportunity to me. An older post can still be valuable and the conversation can take place without the blogger if the community is there, so why shut it down?

Interaction

I also see a lot of interaction (at least on the newer posts) by the blogger. They respond to comments and continue the conversation. This is one of the reasons that they do so well — they relate to their readers and build the sense of community.

Obviously, they can’t respond to all comments on every post. Some comments are not really needing a reply, others are really just restatements of things already said and responded to.

But the A-List bloggers show that they are human and interested in what you have to say by responding to your comments.

Spam in the Comment Box

The final thing that I want to point out in what I’ve seen is that there is very little spam in the comments of these blogs.  I have noticed at times that I get an email that a new comment has been posted on a blog.  Some of these look spammy, but most of the time if you go to the post you’ll find that the comment has been removed quite quickly.  This is a good indication that someone is monitoring and moderating the comments.

The Interview

So, what about the interview questions?  Well, I asked 7 questions (9 if you count the 2-parters as separate questions) and got pretty much what I expected to hear from the respondees.  So, what were the questions?  Here is the list:

1) How many comments do you typically deal with daily?
1b) What percentage is spam?
2) What techniques do you use to filter comments automatically?
3) Do you use "front end" techniques like captcha's to avoid spam?
3b) Why or why not?
4) Do you moderate your comments or just spam comments?
5) What clues do you look for to spot spam that might get through the filters?
6) Do you find much ham (i.e. false positives) when moderating your spam queue?
7) How important are comments to your blog?

Of the 8 major blogs that I contacted, 2 of them responded very quickly and asked me to send the questions over.  They were both generous with their answers so let’s meet them and see what we can learn from them.  A 3rd was a little late to the party, but given the work load he’s under we’ll cut him some slack 8=)

Chris Garrett is a blogger and consultant who helps businesses use the new media tools as effectively as possible.  He’s been around for a long time and written some powerful ebooks that have helped and encouraged me on my journey.  He lives in the U.K. but is planning to rectify that by moving back to Canada where he belongs 8=)  [Welcome home Chris]

Chris Guillibeau is an experienced world traveler and shares his experience and expertise in both the travel and online business niches.  His online empire is called The Art of Non-Conformity and you’d do well to check out his writing to get yourself inspired.  He is based in the U.S. but you’ll find he’s rarely there.  Catch his posts and videocasts from around the globe.

Yaro Starak is the man behind the mega successful Entrepreneur’s Journey.  He has been working online since 1998 and is currently working hard in the Internet Marketing niche, teaching us all how to make a full time income with blogs and membership sites.

[To help separate the answers I’ll use the responder’s last name]

1) How many comments do you typically deal with daily?

Garrett: It varies but across all sites it is probably under 50 comments – it’s hard to say because Disqus and some of my filters mean I don’t have to see the majority

Guillibeau: 200-300

Starak: I personally deal with 10-50 a day, but akismet no doubt deals with a lot more for me.

1a) What percentage is spam?

Garrett: Only a minority now, earlier in the year a group seemed to have targeted me for a denial of service they were sending so many spam messages. Luckily I seem to have come out the other side. At the worst it was hundreds of spam comments a day.

Guillibeau: Maybe 5-10%… I’m not sure exactly. Note that this is just the spam that makes it past the filters. Whatever else that comes, I don’t see but I know it can be quite substantial.

Starak: I’d say 80% are spam if you count what akismet picks up.

2) What techniques do you use to filter comments automatically?

Garrett: I now use Disqus with any comment containing a link set to moderation and I use the blacklist quite heavily. At times I have had to filter at hosting level using IP address also.

Guillibeau: Akismet and related.

Starak: Akismet and in-built wordpress controls for approving comments.

3) Do you use “front end” techniques like captcha’s to avoid spam?

Garrett: No

Guillibeau: No

Starak: No captchas, no.

3a) Why or why not?

Garrett: Disqus seems to be a decent approach right now and I would rather not put additional barriers up

Guillibeau: Don’t want to create a bad or difficult user experience.

Starak: Haven’t needed it.

4) Do you moderate your comments or just spam comments?

Garrett: I used to use moderation for all comments but now just spam

Guillibeau: I moderate.

Starak: I moderate what WordPress moderates based on the rules set up like comments with more than two links are automatically moderated.

5) What clues do you look for to spot spam that might get through the filters?

Garrett: People dropping links or using keywords “SEO Chicago Water Filters” as their name instead of their real human name.

Guillibeau: They are self-promotional, contain links, or are otherwise just irrelevant.

Starak: The biggest clue is what name they use for the comment. If it says like “cheap used forklifts” then I know that’s not a real comment.

6) Do you find much ham (i.e. false positives) when moderating your spam queue?

Garrett: For some reason my brother appears in spam filters a lot, I think because he uses a free online email account.

Guillibeau: A bit, yes — everything mentioned in #1. But it’s not too bad overall.

Starak: There are not too many, at least based on what I know about. Maybe ten a year, though I’m sure there are more than that I never find out about.

7) How important are comments to your blog?

Garrett: Very, although I do understand the arguments from people who have turned them off. My blog is all about building relationships and it’s important for me to hear from my readers.

Guillibeau: Very important at this stage. I don’t have them on every post, but on the regular articles it’s nice to hear from the community. They often add more value than the original post contained.

Starak: Community is important, and sometimes comments offer great additional insights beyond what my blog post provides, so I value them. That being said I suspect my blog would continue to be successful without them.

What Have We Learned?

Based on the responses that I received and the inferences that I was able to make by looking at some of the big blogs out there, the techniques used to battle comment spam by the top guns are pretty similar to what I’ve recommended in this series and they seem to scale reasonably well.

A couple of things to note:

All the respondents made it clear that comments are a vital part of their online work.  They want to hear and interact with their readers in a positive way.  They also made it clear that they want to minimize the hoops that you need to jump through to get there, so no Captcha’s or the like.

They also put a reasonable amount of effort into keeping the spam out using backend techniques to catch obvious spam and moderation to catch the rest.  And it works reasonably well from the sound of it.

An interesting thing that I didn’t see in their answers was the importance of Gravatars in determining spam.  Chris Garrett uses Disqus which doesn’t seem to use Gravatar at all which may be a reason for not seeing that as being important, but the keyword names and unrelated links are something that they all look at.

I hope that you’ve enjoyed learning a bit more about how some of the big guns deal with the problem of spam in blog comments.  What kind of experience have you encountered with comments on these types of blogs?  Do you know of any other A-List bloggers that are doing a great job of balancing conversation and spam fighting?

Related Posts:

• You Deserve Spam
• Battling Comment Spam — A Real Life Example
• Battling Comment Spam — Moderating Comments
• Battling Comment Spam — Dealing with It
• Battling Comment Spam

Tags: a-list bloggers, Chris Garrett, Chris Guillibeua, comment spam, spam