Data is a collection of facts. By itself, data tells no stories. Humans interpret data and craft stories to “explain the data”, hereby providing meaning. This is the heart of data analytics. The digital world has provided the platform where data analytics has driven almost every business, for better or worse.
What are some of the most creative uses of data? Let us understand how data is used, with Facebook.
Taking Advantage of News Algorithms: Facebook
Facebook has, over the years, changed the methods it propagates posts to its readers. Facebook’s profit motive implies that it must remain relevant for its revenue-generating stream, namely advertisers. In other words, Facebook must keep its user base engaged, hereby remaining attractive to advertisers. But how?
Here is where user psychology comes into play. An average user on Facebook would, within the platform:
- browse through the news feed
- follow pages
- read profiles
- use Messenger to share links
A reasonable Facebook algorithm would hence work on providing content that will lead to the user being more engaged on the platform. How do we measure this? We can do this by:
- Counting reactions (like, love, wow, haha, sadness, anger), comments and shares
- Perfoming text analytics on the nature of the comments
- Tracing the flow of shares across Messenger
- Measuring the time taken on Facebook content
- Analysing the clicks the user makes
There are more measurement parameters, but these are some of the more generalised ones.
Algorithms, however, are context-blind, and he who understands algorithms best can ride on it for devastating effect.
The concept behind viral content creation is maximising “engagement”, regardless of the platform. The difference between a physical platform (such as a broadsheet) and a digital platform (such as Facebook) is the feedback mechanism. Most people do not reflect their satisfaction or dissatisfaction through a physical platform; it is too cumbersome to do so. However, feedback through a digital platform is instant. Digital outrage can morph into movements, such as #MeToo. The more controversial the arguments, the more engagement.
Facebook is not immune to us making use of its algorithm, even though we may not know its actual workings. As long as we have a conceptual understanding of the black box, we will be able to take advantage of it and propagate our desired content. Let us illustrate this with an example:
“Hacking” Algorithms
Let us wear the hat of a marketer, who wants to advertise a skincare product. We will want to ride on the algorithm by making sure people would share our content and spend time interacting on it on the Facebook platform. The more traffic there is, the more desirable it is for the marketer. Some simple ideas we will do are the following:
- Creating lucky draws where users are to watch a video (engagement parameter), and answer a simple quiz question through a comment (another engagement parameter)
- Creating lucky draws where users are to hashtag the use of a product (engagement parameter) and tag their friends in it (engagement parameter). When their friends click on the link, they may iteratively continue the sequence.
- Creating lucky draws to share the product with their friends with public share settings (engagement parameter).
None of these are dependent on the actual skincare product — we are merely riding on the parameters Facebook uses to measure engagement. Next, we will organise our social media campaign in the following way:
- Start with a problem statement (dry, flaky and oily skin in harsh weather)
- Show that normal remedies do not work (make the problem statement significant)
- Introduce our product and show “before and after”
- Call to action: do something (e.g. fill a form to book a free appointment with a consultant, or grab a promo code and buy one)
We will also need a payload delivery method. The best method to do so is through third party methods. Hence, marketing companies rely on influencers, and highly encourage community sharing. Who knows the root source of a piece of marketplace gossip?
We have, using a simple everyday marketing example, shown that:
- measurable data are ingredients for algorithms
- the feedback mechanism allows for effective data analytics because data comes back to us
- by understanding algorithms, we can take advantage of them for our own use
But are all data uses this benign? The answer, unfortunately, is no. The next part illustrates a far more powerful use of data, for better or worse.