In a blog last week about Twitter’s new “want to review this tweet?” prompts, I briefly mentioned the “Scunthorpe problem”. I thought this was worth expanding on in a blog of its own, as well as possible ways to get around it.
The Lincolnshire town of Scunthorpe got its name from the Old Norse for “Skuma’s hamlet”, with Skuma believed to be a man once claiming the town as his own. Mentioned in the Domesday Book of 1086, this is long before any residents could have imagined how computers would deal with this name.
In those simpler times, and with English not standardised as we know it today, it also seems unlikely that they were aware of the name containing what we now deem to be a rude word in English. If you’re missing it, this Wikipedia article will literally spell it out for you.
In 1996, AOL blocked Scunthorpe residents from setting up accounts because it believed they were entering profanities into the form. One of Google’s early SafeSearch filters was guilty of a similar misunderstanding, and the aforementioned Wikipedia article even cites that Wikipedia itself took issue with an attempted username of “ScunthorpeM181”.
Traditionally known for its steelmaking industry, the town has now lent its name to the problem of computers treating something perfectly innocent as something offensive, as they lack the knowledge and context to understand that the word “Scunthorpe” is just an English town. Other innocuous words like “wristwatch” and “shitake mushrooms” may be met with similar censorship, and there are reports of people with surnames like Cockburn, Libshitz and Wiener struggling to be taken seriously by computers.
These unfortunate double meanings, and computers’ failing to understand the more innocent uses of them, can lead to problems in email marketing campaigns too. We have recently been working with a client who sells adult toys, and we found that almost all emails we were receiving from them were being filtered as junk by Outlook due to the subject matter. A couple of years ago, I was wondering why my tickets to see the Swedish post-punk band Viagra Boys hadn’t come through to my inbox. They had gone to my junk folder, which wasn’t surprisingly really considering the name of the band and one of the most common subjects of spam emails.
Is there any way around this?
In many cases, not really, which is what makes it so important that sites like Twitter have (hopefully) carried out painstaking research to detect contexts in which seemingly offensive strings of letters are not actually offensive at all. The only real solution if a misunderstanding occurs is to contact the authority in question and try to explain it. I am no computing whizz, but I don’t think it should be difficult to programme a computer to understand that while “Scunthorpe” may contain a naughty word, that word is not so naughty when appearing in the name of a Humberside town.
Where possible, you could use alternative phrasing. Sending out targeted emails promoting “sex toys”, for example, is likely to see you fall foul of countless spam filters, but “adult toys”, although most humans will understand it as roughly the same thing, will more likely be seen by a computer as two “safe” words. In such cases, you can capitalise on computers’ naivety.
In terms of keywords, however, it’s probably more likely that users are searching for the more explicit terms, so you needn’t obscure your website’s content by overusing euphemisms.
Whatever you sell, and however search engines and email filters might view it, we can help you get the most out of it online. Why not talk to Engage Web today – whether you’re based in Scunthorpe, Penistone or even the newly named Austrian town of Fugging.