Image Credit: Mark Knol
Anyone in the UK performing a search in Google on Saturday, between 2.30pm and 3.25pm, would have come across a nasty surprise. Due to “human error”, Google was incorrectly displaying the message “This site may harm your computer” next to almost all search results. Disregarding the warning and clicking on the search result anyway took users to a warning page that told people about the possibility of badware being present on the website they were trying to visit.
Google put out an official explanation here, initially suggesting it was a problem caused by StopBadware.org. StopBadware reponded by saying it wasn’t caused by them, and that it was actually a problem at Google’s end. Google get a list of bad websites from StopBadware, and apparently when Google were inputting the data, they accidentally input “/” – this then flagged every website as badware! Unfortunately by this point the mishap hadn’t gone unnoticed, and people started questioning why Google didn’t have better checks in place to prevent this from happening.
I was out in London on Saturday visiting some friends, and I decided to use Google on my iPhone during this time frame to find a good restaurant. I’d like to think I’m pretty web savvy, but when I saw this message placed against the result I was looking for, it really made me think twice about clicking on it. In fact, it was only when I noticed the same message plastered against every search result that I realised there must be something awry at the Googleplex. This made me think – if I was put off clicking on Google’s results, how many other less savvy internet users would have been scared off. It would be really interesting to see how this effected websites during this timeframe.
Surprisingly, looking through some available data, it doesn’t look like websites were that badly affected. I’ve only been able to look at data for the 24 hour period that was Saturday – i.e. it wasn’t broken down by hour. By comparing traffic data for January 31st against the two previous Saturdays (17th and 24th), it looks like the number of Google referrals remained fairly static. What is interesting, is that doing the same for MSN and Yahoo does show an increase. Referrals from MSN UK appeared to jump by 18%, while referrals from Yahoo UK rose by 14% on what they delivered on the previous two Saturdays.
Is it possible that the biggest damage during this time was to Google’s advertising revenue as users decided to use Yahoo or MSN instead? This research is far from conclusive, so I’d love to hear if anyone else has had the chance to pull some data together on this.
Gregory Lyons on 
Jeremy Head on 
Digital Marketing on 
Ali Ahmadyar on 
Dinesh Thakur on 
Dave Oliver on 
Art Rogue Studio on 
Rob Pearson on 
February 3rd, 2009 at 6:44 pm
I was researching destinations in the Ukraine on Saturday afternoon – and had the same experience. I figured that somehow Google had blacklisted anything to do with the place! Wondered if it was some hotbed for cybercrime or something!
February 4th, 2009 at 4:09 pm
Certainly sent a shiver through the cottge holiday industry! Looking at stats it does seem that google maintained visitor levels compared to the previous 2 saturdays. But direct traffic spikes as did that from yahoo and google. Phoning around it seemed retail was up a huge amount that weekend as well, so i conclude this. Generally buying behaviour spiked and has continued to do so since last saturday. the level playing field google maintained was because there was a spike, and without this i would predict at least 20% drop off. So had it not happened, add 20% where applicable!! Matt evans CEO coastal cottages of Pembrokesire.
February 4th, 2009 at 4:10 pm
that should have read yahoo and MSN!! sorry
February 4th, 2009 at 6:17 pm
Good point Matthew – I didn’t look at any year on year data which might have pointed me to a large spike in online activity on the last Saturday in January. I remember it being freezing cold while I was out on Saturday, so maybe that encouraged people to stay inside and buy online instead, causing the spike you mention.