Economics of digital branding – recommendation algorithms

Not so long ago, I remember when I bought my first car. I methodically jotted down my priorities and all the available choices. After several discussion and debate with my better half, I arrived at the one that best met my requirement. I met a dealer in my area and he made the deal. Though it was quite simple then, the new era of digital sale is not the way it used to be. Does this decision of selecting my car getting transformed in the digital era by algorithms?

Today almost all the consumers have become promiscuous on their brand because of the choices available for them. This is just not the case with the brands that we buy every day but even with the brands that we aspire to buy in future. Today’s customer has the option to connect with myriad brands—through new media channels beyond the manufacturer’s and the retailer’s control or even knowledge—and evaluate a shifting array of them, often expanding the pool before narrowing it. After a purchase these consumers may remain aggressively engaged, publicly promoting or assailing the products they have bought through social media collaborating in the brands’ development, and challenging and shaping their meaning. This constitutes the digital identity of the brand. There are three basic elements to the brand identity. The first one is the value the product or the promise that the product tries to fulfil. The second is the aesthetics or the design element of the product. Finally, the fulfilment, which constitutes the economic model of the entire supply chain. These are the factors that structurally drive the information about a brand on the digital environment. This information with the inputs of recommendations that are provided by the digital media forms the basis of our decision making. In such a case, the choice that we make is our own or driven by someone else?

For example, today we have the youtube channels where a child un-box toys to the delight of toddlers around the world (The Ryan’s toy review) and a Swedish gamer with millions of teenage fans (PewDiePie), running one’s own virtual TV channel with multi-million followers and a few billion cumulative views. Have you ever wondered how they get to the top echelons of viewing? – The trick is done by the internet channels through suggestions or recommendations customised to you. The brain of such channels is – the recommender algorithm. The pioneers of this technique are the early adopters of this technology such as Apple, Facebook, Google and Microsoft which help them to make into the top 10 of the world’s most valuable brands. These are the companies that we touch, speak and feel every day and decide what we consume. If you really think the word of mouth marketing is gone, it is not. Today’s word of mouth is the social media. The key reason why the social media giants such as Facebook and Snapchat are so valuable is just because they are the identities of the new digital human.
The recommendation algorithm is used in a variety of areas including movies, music, news, books, research articles, search queries, social tags, experts, collaborators, jokes, restaurants, garments, financial services, life insurance, romantic partners (online dating), and you name it, it is there. The social media platforms and the e-commerce sites analyse the user’s past behaviour based on the links the user had clicked or the products he had purchased and similar decisions made by other users. This recommendation system can either constitute a top-down approach, in which the broad interests of the user is captured on a regular basis and then compared against the behaviour of other users. The recommendations to the user will not only include his interests but also interests of his peer groups. This is a collaborative approach and is the way google works. Using the huge user base it has google applies recommendations across its platforms. So now you know why once you play a specific video, you get similar videos on youtube.

The other option is a bottoms up approach by identifying the attributes of the product recommended in a standardised form (i.e., the key parameters of search) and identify similar products based on the defined attributes. A nice example is the music genome project. The project gets into the essence of music at the most fundamental level. Over 450 attributes such as gender of lead vocalist, prevalent use of groove, the level of distortion on the instruments, type of background vocals, etc. are used to describe songs. These attributes are mapped to the user preferences to make musical selections of a certain genre based on the user’s preferences. The output is the patented music streaming service powered by the music genome project the Pandora Internet Radio. The movie databases such as Rotten Tomatoes and IMDB also uses the same technique.

Obviously, there has to be a hybrid approach. Netflix uses this approach. It makes recommendations by comparing the watching habits of similar users as well as by offering movies that share characteristics with films that a user has rated highly. This recommendation technique is now extensively used on internet-accessing smart-phones also for GPS navigation and location based services.

But what is the flipside to these recommendation algorithms? They will take off adventure and serendipity on our choices. These algorithms will slowly discourage exploration. Soon there will be a method to the madness and the human behaviour will become more predictable. Amazon will suggest similar books to you, Netflicks will show you the similar movie, SoundCloud will play the similar songs and obviously you will end up taking the same route to the office. So who wins in the end? It would be these internet giants that control the users and their preferences. These goliaths will be able to muscle with the sellers, (I would say the car company, that want me to but the car in future) who want to access the users through their platform. So it is neither me nor my car company that would benefit. It is the broker that wins. This will catapult the oligopoly of our internet companies that touch our daily lives to a different level – driving them to not billion dollar enterprises but trillion-dollar enterprises.