Some of you may have noticed some delay in posting last week.
The cause of this, briefly, was that Google decided that my blog was a suspected splog (spam blog). For those of you who've managed to avoid splogs so far, these are automatically produced blogs that get their content by more or less randomly cutting and pasting pieces from other sites on the web. The effect is rather strange - at first glance, the splog looks like a legitimate blog and each small fragment likewise looks legitimate, but after 10 seconds of looking you generally realize that the blog as a whole makes no sense. I believe the reason people make splogs is to generate traffic from search engines or other illegitimately obtained links, and thereby make money from ads.
This note may seem a bit ungracious, considering that Google gives out lots of cool services for free and splogs are a growing problem that they do have to tackle. In fact, I wasn't going to write this until the blog my girlfriend is developing was also flagged. This made me think some suggestions to Google might be in order, specifically about a couple of idiocies that they should probably think better of.
The first idiocy is that for this, as for their other services, Google makes it impossible to contact a human to fix problems; the only message from Google was a no-reply email with a notice that my blog had been locked to new posts and would be deleted in 20 days, and a link for me to click if I believed my blog wasn't spam. if I did this it was intimated that Google would get around to checking things out sometime in the next 2 business days and unlock the blog if they did decide it wasn't spam. Let me suggest to Google that they change this no-human-contact policy since it's really dumb. I don't have much hope this suggestion will be taken because it does appear to be a fundamental principle at Google that the algorithm is to be trusted over human judgement.
Second, If you do insist on giving algorithms primacy over humans, it would at least behoove you not to use lamebrained algorithms, as you did in this case. A few examples of how a smarter algorithm could easily tell this blog isn't spam:
- It has been mentioned on and linked to by writers for the Financial Times, Wall St. Journal, The Atlantic, The NY Times, and most recently by the estimable Australian economist Steve Keen (that website has lots of interesting stuff by the way). The odds of these people talking about and linking to a spam blog are close to nil.
- It has around a thousand subscribers from many countries who read items regularly; while in theory this could be faked, I'd guess very few splogs take the trouble, so it might make sense if you find this situation to delegate a human to spend the 10 seconds necessary to check if the suspect blog is in fact spam.
- My girlfriend hadn't launched her blog yet, so it was set to be invisible to search engines and other readers. The odds of a spam blog having this setting seem very small. I'm guessing the spam detection was set off because she had a bunch of repeated text in dummy blog posts, but this seems it would like a fairly common thing for authors who are setting up the design of a new blog.
Once again, I'm sorry to have to complain like this about a free service. I'm sure your spam detection algorithm was programmed by a crack team of dual linguistics-computer science Ph.D.'s and is the last word in sophistication, but I think a little common sense might also be added to the mix. While I'll continue to use your blog service since it's free and pretty good, this experience has certainly reduced the possibility of my using a Google service for anything crucial to my financial or physical well-being without some policy changes on your part.
Thanks for sharing this. Hopefully Google will read it and it will make someone there realize that their s--- does indeed stink, much like everyone else's.
ReplyDelete