South by South West Interactive (SxSW) is no ordinary conference. It is a festival and a platform that allows for the showcasing of content fused with technology.
Here, robots meet their designer, the social dimension meets analytics, and Snowden meets the rest of us. In many panels, content and big data are discussed in the same breath. Until recently, content creation – including blogging and tweeting – were completely separated from data analytics.
The unique qualities of a writer have traditionally been an impeccable command of language and unlimited creativity, while those of an analyst have been to write SQL code or run a logarithmic regression.
At SxSW, discussions on those skills have merged. Is this a lack of differentiation? No. It is an indication that this industry began to come of age 10 years after the Facebook craze took off.
We are in the age of the algorithm. An algorithm is the glue that binds content, data and insights.
We have an unprecedented amount of digital content from which the SxSW is breaking records and making headlines.
For instance, my former startup Fisheye Analytics (sold to WPP a couple of months ago) counted 14 thousand articles – blog and news posts – and close to 100 thousand tweet counts on the subject of SxSW.
By any standard, that is a lot of content. It is safe to assume that no one will be able to read all it all. But is there a way of sifting through all the information and understanding what is important? Yes. How? Well you guessed it correctly: through an algorithm. Algorithms can measure the number of shares and engagements.
Bjoern Ognibeni (@ognibeni), co-founder of BuzzRank, a social media analytics company, pointed out in a brief video interview that if those metrics are incorporated into an algorithm, they can be used for curation.
Many successful startups have grown in this space, including BuzzRank with its Curator and tame.it. Big players like LinkedIn, Facebook and Twitter created their own stand-alone algorithms to supply their audience with relevant content. (To read more about the metrics powering those algorithms, click here).
Algorithms are the cure for the “content chaos,” and they are here to stay. Steve Rosenbaum (@magnify) explained in his talk at SxSW that a wise combination of human judgement enabled by algorithms will become the new king of content.
If you missed those discussions on SxSW, I recommend you read through some suggestions made by one of the masterminds in curation: Robin Good (@robingood).
BUSINESS DRIVEN ALGORITHM
Does automated curation pose a potential problem to the already troubled media industry? I don’t think so (and have argued so much at many conferences – here is my talk at Ria Novosti). The media industry could make money by using curation metrics quite effectively.
Let’s start with the tabloid press. They have always been eager to get viral content up and running. To use the word ‘viral’ is actually wrong; we could more accurately describe such content as ‘contagious’ (learn more about why that is).
Engagement metrics and curation algorithms are great tools to find contagious content. An example? Yes, cat videos. Don’t laugh! At SxSW people were lining up to meet grumpy cat. She appeared on stage together with Ben Lashes (@BenLashes), the world’s first meme manager. Ben very vividly described how people can make a viable business out of cat content.
But the usage of data metrics is not only limited to the tabloid press. Quality media houses and journalists can also benefit greatly from algorithms. Scott Havens (@msh200) from The Atlantic argued that many media outlets with good reputations failed because they lacked data on what their audience wanted.
Marshall Kirkpatrick (@marshallk), Founder of Little Bird, explains in this video that there are other ways in which the media industry can utilize content data beyond curation. He has built a system to find relevant experts for any kind of topic so that media companies can react quickly to news and engage the right expert at the right time.
Marshall’s algorithms are not only focused on finding important content, but also allow one to find important people. I deliberately do not call important people “influencers” because research has shown that they do not exist.
Personal identity is an asset and personal identity is data. This is known to 277 million registered LinkedIn users, as well as to a whole class of startups that try to market and sell to audiences based on profile behavior.
However, whenever an algorithm is successful, there will be questions on how to skew or tweak it.
For example, Will Lowe (@willow9886), in a workshop dubbed “Hacking LinkedIn,” shared the tweaks that he has made to influence LinkedIn’s algorithms to make his CEO’s profile appear more often in searches.
Algorithms are powerful. They help us to find our way through the chaos; they show us who is important and who is not.
“These little pieces of code are more powerful now than a lot of the most powerful editors in media,” said Eli Pariser (@elipariser) during a podium discussion.
Eli’s statement brings forth an underlying issue. The more powerful some code is, the higher the likelihood of unexpected consequences.
On one side, there is an incentive to skew an algorithm. Google’s search algorithm is the best one to use as an example.
How Google ranks websites can make or break businesses. This has led to the rise of experts that carry out Search Engine Optimization (SEO). Businesses are paying these experts top dollar so that they influence Google’s algorithms to favor their sites. Google reacts by changing their algorithm regularly to keep it a secret.
On the other side, this very obscurity means that those extremely powerful algorithms are not governed by any public control. Gilad Lotan (@gilgul) from the startup incubator BetaWorks and Kelly McBride (@kellymcb) from The Poynter Institute put this into perspective in a very engaging podium discussion.
The #occupywallstreet hashtag did not trend on twitte,r but Kim Kardashian’s wedding did. Why? Definitely not because the movement was of less importance in comparison to the wedding. The algorithm was designed in a way that it only found bursty conversations, resulting from situations where a lot of people tweeted within the same time frame.
Algorithms are not neutral. They need to make decisions in favor of one behavior over another. They determine what is trending and what is not. Because they are so powerful, they have a bias that determines which information gets heard or which person is important. As Gilad points out in this video interview, so far, there is no one to control this power.
LEARN THE TECHNOLOGY
We are in the era of the algorithm. They decide what news we will see, they decide which person is important and they will continue to merge into our non-digital lives.
Most of those developments are good news for all of us. We all love Google for its algorithm, and we go to curation sites for their algorithms. But as algorithms rise in power, there is a danger as well. The danger is not in the fact that those algorithms are obscure. The danger is that any algorithm might fall prey to someone trying to influence it. This might be the ones programming the algorithm or the users. We for instance saw governments trying to skew algorithms by introducing fake online personas (Learn more about the US government persona-management software).
But the biggest and realest danger lies in us. If we believe that there is only one truth and that is the one generated by a black-box algorithm, we might be deceived easily.
Each and every one of us needs to learn the basics of how those algorithms work. Subscribe to my newsletter (here) and I will keep you regularly updated on wrong and misleading metrics and skewed algorithms.