Fooling The Ephemeral Internet, One Trick At A Time
By Prateek Waghre:
The fact that a number of black hat tricksters have been around trying to game the internet is a well known fact. People using ‘crooked’ practices to boost their search engine rankings is one of the reasons Google cites for keeping its search algorithm such a closely guarded secret.
But those are black hatters. What happens when average joe’s are able to outsmart the system without doing anything radically ‘Black Hat’. In fact, some of them don’t even qualify as White Hat. Over the last 7-10 days I’ve stumbled across 3 such instances, from tactics similar to flash mobs to good old fashioned repetition, people have been trying to find ways to increase their relevance period in the increasingly ephemeral internet.
Tweet More to Score More!
The first instance concerns the weapon of choice for many – Twitter. Adrian Pelzer set out to study whether merely tweeting a lot is a way to gain popularity on Twitter, and increase your Klout score. I realize Klout is by no means a standard for measuring influence on Twitter, still many people tend to regard Klout scores with some legitimacy. Twitter apps like HootSuite tie it into profile displays too. He set up 4 bots that sent out tweets at fixed intervals – once in a minute, once in 5 minutes, once in fifteen minutes and once in thirty minutes. They were anonymous, no bio, no avatars and did not follow anyone.
He let them run for 80 days after which he compared the number of followers and their klout scores. The gist of it was – The more you tweet, the more followers you have and the higher your Klout score is. The quality of tweets din’t matter as much as the quantity. Of course a percentage of the followers were other bots. While PeerIndex, though it kept the rating of the bots low, gave these bots a 75% realness score, which meant that there was a 75% chance it was a human! It gave some known human users a significantly lower score.
Now you see how influential spam can be on twitter! And all those people who are constantly on the lookout for seemingly ‘intellectual’ tweets to increase their follower count – you might find this useful.
Open May not be so great after all
After receiving plenty of bad reviews, these rogue developers would pull down their apps from the app store and re submit it to the app store with a new package name. And then give it a 5 star rating as soon as it went back on. This practice also ensured that the app stayed on the ‘Newest’ list and that it got picked up by sites like AppBrain. And when users voted it down again, they would repeat the whole process. The post cited that one of the apps had appeared 175 times. Somewhere I can hear Steve Jobs having a hearty laugh.
Most Read News?
Now, if like me, you are someone who does most of your reading, news consumption online. This next example is unlikely to make you happy. It isn’t possible for any human to pick out the most relevant and popular stories all the time. Which is why several sites have incorporated filters like Most Viewed/Shared/Trending etc. Daily Beast’s Thomas Weber set out to find out how many emails it takes to break into New York Times’ Most Emailed list. The NYTimes is one of the most visited news sites on the web with a US Alexa rank of 27 and receives about 50 million unique visitors each month. And unlike a lot of other websites, that display their most read stories, NYTimes showcases its most emailed stories far more prominently, because it believes that mailing a story is a greater sign of engagement. We’ll buy that for the time being.
What Thomas Weber did was simple. Similar to a real world flash mob, volunteers would descend onto the New York Times during a certain period of time and email a pre selected, obscure and uninteresting article – that someone else was unlikely to share – to other people.
First they determined that it was the number of users who emailed it that mattered, and not the number of people it was shared with. It took under 50 senders to break into the ‘Popular’ stories in the science section. But obviously needed a significantly higher number to reach the most emailed list on the front page of the website. It took all of 1270 mails to break into the top 5 Emailed stories – over a period of 24 hours. Which, if you look at it in terms of its actual traffic, really isn’t that high a number. And sure, some might argue that Most Emailed isn’t as indicative as Most Viewed. But at the time of writing this post – 8 of the Most Emailed stories appeared on the Most Viewed list as well. So there is definitely a strong correlation between the lists. Besides, an article in the most emailed list, is going to attract plenty of page views by virtue of being there.
On all accounts, it seems like this experiment was done outside of peak traffic hours on the website. But isn’t the number of emails that it took, for the story to break into to top 5. It was the relative ease with which the entire thing could be orchestrated that is cause for concern.
I’ve also previously written about the Resig brothers, who routinely make viral memes with the aim of seeing them rise in popularity. And while, two of these three examples were investigative in nature, the ease with which they were carried out shows us that some of the filters we use on the net are very susceptible to being ‘gamed’ and I can think of countless ways to use them unethically for profit. As I’m sure many already have. And some are probably doing too. There is serious cause for concern as more of our world moves online, where most tweets are lost within 30 minutes of being posted and we are constantly dealing with a barrage of information.