A machine-learning algorithm walks into a bar. The bartender asks: “what’ll you have?"....
The algorithm replies: “what’s everybody else having?”...
How Algorithms Affect Just About Everything
Do you have any idea how widespread the effect of algorithms is on our everyday life?
Whether we’re building a Lego toy for the kids, or assembling a new IKEA sofa, simple instructional algorithms can (usually) be understood and carried out by us mere humans, compared to those algorithms which are embedded in code, and affect everything from online grocery shopping to the facial recognition software used by law enforcement agencies.
As another example, it's thanks to Google’s PageRank algorithm that we get the search results we want in fractions of a second, which helps us to make sense of the vast resources available to us on the internet.
If you’ve ever been reminded by Ocado that you’ve forgotten to order avocados this week, it’s all thanks to the humble algorithm. But that’s just the tip of the iceberg. They might also suggest complementary products (wholegrain bread, craft ale, anyone?) based on other users’ preferences. And we’ve all had Netflix or Amazon suggest what ‘you might also enjoy’. The modern algorithm machine is pretty sophisticated.
But how does this affect digital advertising? One of the most significant developments in this field has been that of ‘programmatic buying’. The profiling of internet users based on their web activity isn’t news, but in terms of advertising, this profiling means that we can now predict – with some accuracy – the likelihood of a user’s interest in a specific ad at any given time. In other words, algorithms have made it possible to calculate the precise value of online advertising space, on an individual user basis. Pretty cool.
But with great power, comes great responsibility, and we also need to think about the importance of ethics; of operating responsibly within this developing system and maintaining a good ‘e-reputation’. Enter, GDPR
. Rightly, this new data protection legislation will make sure that the people handling user information comply with personal data protection standards, beginning at the design stage (aka ‘privacy by design’). No doubt GDPR will, in time, bring with it further regulation of the algorithmic processing of data, but demands for greater transparency in the industry are both inevitable and welcome. If modern tech allows the digital marketing industry to influence consumer choices – whether overtly or subliminally – then there needs to be accountability in that domain.
There’s also been some heated debate about the lack of neutrality in algorithms. Twitter hit the headlines recently
, when the company’s chief executive Jack Dorsey told US lawmakers that Twitter’s algorithms had not always been “impartial”, and admitted the company had deliberately reduced the visibility of 600,000 accounts, including members of US Congress. Facebook’s latest algorithm update – implemented in January this year - now puts posts “that spark conversations and meaningful interactions between people” at the top of your News Feed, in other words content which receives the most engagement - reactions, comments and shares - will appear first, which has made things considerably harder for digital advertisers. In contrast, Instagram has reverted to its original ‘recency algorithm’ as a direct result of pressure from users. Twitter shows you ‘relevant tweets’ first, based on your interaction history with the accounts you follow, and LinkedIn’s complex algorithm relies on a variety of metrics related to your personal profile, content engagement and your previous behaviours, to prioritise the posts you see.
One thing’s for sure: algorithms are here to stay, and as computer tech advances at a rapid rate, so does the complexity and sophistication of the algorithms we’re exposed to on a daily basis. Most of these aren’t even written by humans anymore, but – you guessed it – by other algorithms. The implications for tailored marketing are almost limitless, but as digital marketers, we need to focus on the most effective, customer-driven applications of algorithms, those that help both brands and consumers alike. Let’s not forget: marketing has always been about understanding people, and that understanding has always been driven by the collection and analysis of data. It’s just that now, we’re getting a little bit more help from technology.