Earlier we brought you news of two prominent travel bloggers subpoenaed and threatened by the TSA. Here is another story of a prominent blogger under attack this time by Google’s spam robots. John Hempton has a very popular financial blog, Bronte Capital, that he’s published on Blogger since May of 2008. As most of us know, Blogger is owned by Google.
Google shut down John Hempton’s popular financial blog after Hempton wrote a detailed and controversial post titled: A dark privatised social security story: Astarra, the missing money and how examining a fund manager owned by Joe Biden’s family led to substantial regulatory action in Australia
Felix Salmon, a financial journalist who blogs for Reuters, broke the story on his blog:
On Saturday, January 2, John Hempton put up a monster 3,200-word blog entry about a fund group in Australia which has all the appearances of being, he wrote, “qualitatively different and more serious than any previous fund collapse in Australia”. He was featured on the front page of the Sydney Morning Herald (other SMH stories on this subject are here and here), and his blog was read around the world, partly because it continued his investigation into Paradigm, a fund-management firm owned by Joe Biden’s son and brother — a company which, he wrote, “keep being associated with cases like this”.
At the time Salmon wrote it, his theory was that someone who wanted to shut the blog down reported it as spam.
Here are two screen shots Salmon posted on his blog:
Here is a screen shot from Hempton’s perspective:
It does seem to be a case of very odd timing and very stupid spam-prevention robots. As Salmon argues, Hempton’s blog is long-running, well-trafficked and has tons of incoming links from various sources including the Wall Street Journal, Reuters and the Financial Times. You’d think a bot would be able to figure it out or, at least, that Google would have a mechanism in place to flag a real person to double-check the blog before preemptively shutting it down.
Hempton’s blog is back up now because Google’s Rick Klau saw Salmon’s post and unblocked Hempton’s blog. Here is what Klau said about this problem:
We don’t go into great detail about what triggers a takedown, for the very reason you mention: we don’t want to make it easy for spammers to game the system. That said, we also have a number of protections to avoid precisely this situation – before I document what those are, I’ve asked the engineers responsible for maintaining them to explain to me why they didn’t work. Generally speaking, there were a number of indicators for John’s blog that should have very easily avoided any false positive spam classification – and I’m trying to find out how/why those got bypassed.
Rest assured that this is an internal bug – not any external players gaming our abuse reporting to try and take down John’s blog.
John – for restoring access, we actually met just ahead of the holidays to go through several ideas about making the resumption of service for legit blogs caught in a false positive situation (like yours) more seamless. Your idea is quite similar to reCAPTCHA (http://recaptcha.net/) – a service Google purchased in 2009, and one which we will be incorporating into Blogger in Q1. Glad we’re thinking along the same lines.
It’s good that it’s all sorted out, and it’s also good that Google is working on solving the problem. A blog like Bronte Capital ought not be shut down without a very good reason.
Google “production” software is notoriously full of bugs. They appear and disappear with great regularity, something that frequent Google users come to know. I use Google/Books a lot. One night I enter a search key that generated about forty pages of results. As I was working my way through this list, I got to page 21. The WEB-site denied me access to any pages past 20–claiming that I was a “crawler” (or some such). I reported this problem, and a few weeks later it disappeared.
This sounds like bad timing .. that’s all.
I agree, at it does sound like bad timing. I’d say what put him on the express track was that Salmon wrote about it and that caught someone’s eye at Google. It was a bad case of odd timing.
I have a Blogger blog and when writing this I thought of the day Google bots took my little blog out. I’m certainly not as influential as Hempton, but I reported it and Google was pretty fast with reversing it. Like him, in my niche, I had a fair amount of links, a long time on the web and the blog was active.
It just seems that there should be certain criteria that get some blogs kicked to a real person before they get shut down. However, who knows if Hempton’s blog would have made that cut.
It shows there’s one rule for those with friends (or at least followers) like Felix Salmon, and one rule for the little guy. I’ve been fighting Google on getting a business blog restored for a friend since July (a taste of that story is at http://www.infonews.co.nz/news.cfm?l=1&t=0&id=46498). Right now, I’m feeling rather envious of Mr Hempton, who got his blog back in a day, while the “official channels” have been obstructive and a tad thick.