Tuesday, November 28, 2017

You'll Never Guess What's Ruining My Metrics

I delete anywhere from five to fifteen subscribers per day from my blog. If I didn't do this, I would have well over a thousand compared to the fifty I have. Why do I do this when begging for subscription is part of this new age, social media, selfie driven, narcissistic world in which we live in? Because they're all fake, that's why.

A few months ago, I noticed a spike in subscribers which immediately had me thinking that I crossed an unseen threshold in viral interest in stories about celebs, politics and diarrhea. At closer inspection, I noticed all the emails were from outlook and had odd, exorbitantly long usernames. Who would create such an encumbered email? The bogus subscribers look like this. 

Bogus Emails
No one intentionally would type "mascaraquettmuriellmo" every time they wanted to check email. I searched the internet in attempts to uncover what's going on. In a forum, I learned that this infection of fraudulent outlook emails with DNA like usernames is called "Subscription Email Spam."

Bots, computer programs that surf the web, look for blogs like mine to skew analytics such as bounce rate and session duration. They also consume my advertising budget by programmatically clicking on my ads. Apparently, they do this because someone is unhappy with something I wrote. What's puzzling is that to subscribe to most blogs you got to get past a CAPTCHA, the completely automated, public Turing test to tell computers and humans apart. You know, the annoying graphics which pose a masked series of characters that a human can read, but a bot struggles with. This is a typical CAPTCHA offered by the Google subscription service, Feedburner.

Pretty straight forward. Some CAPTCHAs are more difficult to decipher. Once I was presented with this when being forced to register while buying a toaster.


No one needs toast that bad. Also, for some unknown reason I really wanted a unicorn. CAPTCHAs are annoying, but they're affective at warding off malware written by smart people who are probably fans of Star Trek. So the obvious question is how does a bot programmatically get around a CAPTCHA? The answer is: porn.

A Google search revealed that apparently some clever developer wrote software that copies the CAPTCHA image then reproduces it on a website that offers nude picture if one enters the correct sequence of characters. This allows a bot to match a string of characters to an image which then can be recalled later to circumvent the Turing test. It's hard for me to believe with all the porn on the internet that there are humans so desperate that they will bother to decode a CAPTCHA in order to gain access to even more porn. Let's face it, there's no shortage of nude people on the internet.

Back in the early 1990's when the internet became more widely accessible to the public, there wasn't a lot of information out there. The first time I searched the internet, I queried on "Abraham Lincoln" which returned four hits. Back then, there wasn't a lot of porn out there either. I recall the first time I saw a website which hosted naked pictures of humans. The web developer recorded who visited the webpage and for how long and presented a list of all the users who most frequented the site. The list was entitled, "The Internet's Horniest Geeks."


Kim 
Porn can have unintended consequences. Kim Kardashian once tried to "break the internet" with a picture of her bare ass. She presumed hordes of horny nerds, frantically downloading a picture of her bulbous butt, would crash the internet. It didn't work. The internet didn't grind to halt. I think if Kim wanted to bring down the internet, she would've had a better go of it if she sat on it.

Thanks indirectly to the most jacked up dorks among us, I've got scores of bogus subscriptions flooding my blog daily because someone is mad at me for poking fun at Stephen Hawking, Hillary Clinton or Kim Jong un. Luckily, I still remember enough web programming from my career as a computer scientist to write a script to purge suspect emails for me. I used to write software before my job was offloaded overseas to a country with marginal indoor plumbing as a cost saving measure recommended by financial bean counters. Unfortunately, I'm not smart enough to figure out how to prevent the circumvention of CAPTCHAs via porn. I would suspect that the solution lies somewhere in randomly changing the CAPTCHA image identification since its doubtful that the work of the bots occur in real time.

I guess in the end it comes down to the fact that one should never underestimate the elegant programming of a lonely geek.

No comments:

Post a Comment