For some time the Askimet plugin/service has been the go-to spam filter for WordPress websites. Although Askimet is good in a lot of ways, it’s not without its problems.
Askimet is a ‘learning’ spam filter which after a given amount of time makes its own decisions on what constitutes spam and which visitors to block comments from. While it works well in many instances, it does make mistakes and treat genuine visitors as spam.
To be fair, some margin of error is to be expected, and I can’t criticise Askimet for not being 100% perfect. Where it does deserve criticism is in the way it handles mistakes.
Ideally, comments flagged as spam would go into your spam comment queue in WordPress (which many do) so that you can review and retrieve any legitimate comments mistakenly flagged as spam. Askimet can also automatically delete spam posts older than one month, a feature which also becomes a must once you’re receiving hundreds of spam comments a day or more.
The problem is that instead of the above happening where comments can be retrieved, users leaving legitimate comments which are incorrectly flagged are presented with the usual “Your comment is awaiting moderation” message, which would suggest that leaving their comment has been successful, however the comment is automatically ignored/deleted and the blog author is never aware that an attempt was made to leave a comment.
The problems with this are obvious. The website author is missing out on comments, a potentially valuable resource for user generated content and website community tool, and the user is left disappointed that the website author has seemingly decided to delete their thoughtful comment.
Note: It is possible to see if the above has happened at the time of posting. When a comment is successfully published, WordPress takes the user to the comment URL of their comment. i.e. http://www.andrewkeir.com/post-url/#comment-1234. When a comment is not accepted the “Your comment is awaiting moderation” is shown, but the #comment-1234 portion of the URL is not.
Whatever small peace of mind it may be for visitors to know that they have been incorrectly flagged and not ignored, it still doesn’t solve the problem, and in any case, your visitors can’t be expected to know the above information.
Even worse, some instances of Askimet return a white screen when the ‘submit’ button is clicked to publish a comment, suggesting that your website is broken. Some months ago I was prevented from commenting on my own blog, and presented with a white screen when attempting to reply to visitor comments. Don’t ask me how I get treated as a spammer on my own website…
There are alternative to Askimets like CAPTCHA text, but these too have some usability issues. I think it’s safe to say that everyone has entered capture text incorrectly on more than one occasion. I was recently presented with a CAPTCHA text prompt that somehow showed a word and a picture of… something? A door, a vent, wood? I don’t know what I was supposed to type here and was forced to refresh and start my process again.
GASP (GrowMap Anti-Spambot Plugin) works differently by presenting a simple tick box that is not visible to bots, and doesn’t have the illegibility issues that CAPTCHA often presents. By not being visible to bots, you eliminate the vast majority of spam comments before they can even reach you, while removing the potential to have legitimate comments disappear without your knowledge.
It is still possible for human spammers to manually leave a comment, ticking the GASP box and sending it into your moderation queue, but in my experience this is much more desirable than the alternative.
While using GASP, I have been receiving a dozen or so spam comments a day. These are immediately identifiable and can be deleted in a few seconds. I also don’t have to worry about comments being blocked without my knowledge.
While using Askimet I was receiving up to 500 or more spam comments a day. I had to scan through pages of spam each day to check for any mistakenly flagged comments (which did come up from time to time) and was no doubt missing others which weren’t even making it to the queue. As your website grows so will your spam comments, and this just becomes more and more unmanageable.
Which spam filter do you use?