Customer experience personalization is all about data first. Get the data right and you can shape the overall customer experienceOpens a new window by applying data science and machine learning. Recommendation engines are very powerful personalization tools because it’s a great way to do “discovery†– showing people items they will like but are unlikely to discover by themselves. They improve a visitor’s experience by offering relevant items at the right time and on the right page. In the immortal words of Steve Jobs – “a lot of times, people don’t know what they want until you show it to them.â€
Because of how well recommendation engines boost subscriber numbers through engagement and stickiness, facilitating such serendipitous discovery has turned into a high stakes multi-billion-dollar race for the world’s biggest digital companies. Personalized suggestions are implemented via software programs that crunch massive amounts of data to “learn†user preferences and come up with a list of recommended items for the user.
The customer personalizationOpens a new window journeys of Amazon and Netflix demonstrate just how powerful recommendation engines can be. See how these online giants built cutting edge recommendation engines that keep subscribers coming back for more.
Amazon
( The image describes the recommendations across the buying experience — from product discovery to checkout)
A lot of Amazon’s fantastic revenue growth has been built on successfully integrating recommendations across the buying experience — from product discovery to checkout. Enabling personalized suggestions in e-commerce, is perhaps the number one reason for recommendation engines, because of what is known as the long tail problem – rare, obscure items that are not very popular and don’t drive the bulk of revenue. Recommending long tail items to shoppers is critical because if successful it has the potential of giving ROI on slow-moving inventory.
The retail giant’s recommendation algorithms are based on seemingly few elements: a user’s purchase history, items in their shopping cart, items they’ve rated and liked, and what other customers have viewed and purchased. However, for a retailer with as many items as Amazon, the challenge becomes which recommendations to present and in which order – a problem known as “learning to rank†in data science. A secondary problem is one of diversity – how to show a diverse selection of items in your recommendation. Amazon’s is able to achieve a high level of customer relevance though algorithms based on a process called item-to-item collaborative filtering.
Learn More:Â What is Customer Experience (CX)? Definition, Design, Management, Best Practices and ExamplesOpens a new window
The importance of suggesting the right item to the right user can be gauged by the fact that 35% of all sales are estimated to be generated by the recommendation engine. Amazon is investing a large amount of talent and resources on getting better to AI – specifically “deep learning†technology to make recommendation engines which learn and scale even more efficiently. Deep learning involves massively networked computing power to enable more complex forms of machine learning.
In May 2016, Amazon opened up its sophisticated AI technology as a cloud platform. The company unveiled DSSTNE (pronounced “destinyâ€), an open-source artificial intelligence framework that Amazon developed to power its own product recommendation system. “We are releasing DSSTNE as open source software so that the promise of deep learning can extend beyond speech and object recognition to other areas such as search and recommendations,†said the company.
It’s easy to understand why Amazon has a strong incentive for open sourcing its AI, thus increasing the probability that a developer outside Amazon will find a way to make the recommendation system better. Ultimately, the retail giant wants to create a system that can make get better at predicting products based on lesser data and of course products that customers are more likely to click on and buy.
Netflix
( The images showcases the highlighted content library on the Netflix Homepage)
No one understands the idea of content discovery better than Netflix, because the on-demand streaming video is probably the world’s biggest market for digital consumption of content. Netflix has worked hard to ensure its recommendation algorithms can highlight as much of its content library as possible. The savings produced by the Netflix algorithm, show up through increased viewership and lower churn. Netflix’s recommendation engine tuned for hyper-specific categorization can match tiles to the exact people who would be interested in them.
Strong recommendations also increase the number of time viewers watch content on Netflix keeping subscriber churn as low as possible. According to a paperOpens a new window  (Click here to read about various algorithms that make up the Netflix recommender system, the role of search and related algorithms) published by Netflix executives, the on-demand video streaming service claims its AI assisted recommendation system saves the company $1 billion per year. This means Netflix can confidently spend huge sums ($6 billion a year) on new content, knowing viewers will consume enough overtime to give them healthy returns on the investment.
Realizing the importance of having the best recommendation engine, Netflix puts a lot of effort into optimizing its algorithm. Updates to the algorithms are researched and tested by a team of over 70 engineers. In 2009, Netflix offered a $1 million prize in an open competition to any research team which could improve on the efficiency of their algorithms. The Netflix Prize was an important event in the development of content discovery systems — shining a light on recommendation engine technology, and bringing new machine learning scientists to the topic.
Last December, Netflix again revamped the technology behind its content recommendation engine – deciding to do away with region-based preferences in light of their ongoing global expansion. A decision was taken because of what is known as the “cold start†problem – a limitation of such systems. To work well, the software needs massive data sets to crunch. In a region-based approach to personalized suggestions, Netflix faced this issue with every new territory launch – with this update, Netflix can start to scale the engine on a global level.
The next challenge is coping with different languages, but the best part about their investment in this technology will be that, as the user base increases, Netflix’s recommendation engine only gets better. More viewer data strengthens the algorithm and results in more insights about viewer behavior. Clearly, Netflix is aware of what a huge weapon personalized content discovery can be, achieve its global ambitions.