The internet seems to be growing more and more intuitive. Have you noticed? Most of us have had the experience of combing the web for the best price on that new gadget we’re dying to buy. And when it’s time to get back to work (it’s ok, we won’t tell), an ad for the very same gadget is stalking us from website to website! Below we will explore the basics of this type of advertising, known as retargeting, and also dip our toes into the pretty science-y stuff that has developed since. Well try and answer the question, how does the internet know me so well?
What Is Retargeting?
Another type of retargeting is list-based, where people have given their email address in a previous interaction. The list of email addresses can then be uploaded to a retargeting campaign, using platforms such as Facebook or Twitter, and the platform will identify users with those email addresses. This makes it very easy to target only those users and serve up ads appropriate to them specifically.
But what about those times when the internet takes it a few steps further and shows you a relevant ad on your phone when you have only searched for that ironic t-shirt from your desktop computer? Or when it begins to identify the faces in your photos, or tells you that you may need to take an alternate route to work because of heavy traffic? This is a step beyond retargeting and is truly remarkable technology.
Neural Networks and Deep Learning
Think about the amazement you feel the very first time Facebook suggests you tag a photo of a person it has identified correctly as your sister, or the slight discomfort and simultaneous feeling of awe when Google Photos makes a collage featuring your new puppy. These uneasy feelings, tinged with glee that the capability exists, are all thanks to technologies called neural networks and deep learning.
Simply put, Neural Networks are “machine learning models…[based] loosely on the network of neurons in the human brain.” The technology involves far reaching mathematical models that analyze incredible amounts of data to actually learn. Neural nets have completely reinvented photo recognition, speech recognition, and speech translation, “the ability to automatically translate speech from one language to another.” Perhaps the most impressive characteristic of this machine learning is that even when provided with flawed data, it can learn beyond those flaws if given enough data.
Deep Learning works by continuously recycling data. Algorithms interpret raw data and use that output as input to create a more abstract representation of it. “As a result, the more data is being fed into the right algorithm…the apter it gets at handling new, similar situations.” Read: it learns. Because of this incredible ability to learn, Deep Learning models will undoubtedly be utilized by eCommerce stores and marketers alike to deepen personalization. Personalization has been proven to drive sales and to take the user experience to a higher level.
The data collection used by Google alone amounts to an impressive overview of your life. Your driving habits; your shopping preferences; the things you search for and buy. Google knows it all. Utilizing neural networks and the process of deep learning, Google (and other platforms) aims to enhance your experience of the Internet. But as there are always pros and cons with topics that involve privacy, Google also provides ways for you to limit what kind of data is collected.
These revelations can be a bit frightening, or feel invasive. There is undoubtedly a line between enhancing user experience and seriously invading one’s privacy. There is no denying, however, that the fact that these technologies exist – that an internet platform can learn – is pretty impressive. So go on, personalize your Google account and all of your web browsers. Determine how intuitive you want your internet experience to be. When it comes to asking, “How does the internet know me so well?” Know that you are in control.