‘People you may know’, ‘other movies you may enjoy’, ‘jobs you may be interested in’, any of these sound familiar? If you use the internet frequently, you are sure to have come across suggestions like these. What you may not know though is that these suggestions are made with the help of a recommendation engine. So let’s see what exactly these recommendations are and how UX is redefining them.
Recommendation engines – What they are and how they work?
Recommendation engines or a recommender system is software that analyzes data to provide the user with suggestions. These are frequently used all over the internet. Be it for online shopping, finding a job, a partner or even while searching for information.
The software use complex calculations to make sense of vast amounts of data. They mathematically calculate people’s actions in relation to one and other. Broadly, a recommender system will work either with a content-based approach or a collaborative approach.
A content-based approach generates suggestions based on the user’s previous choices. For example, if you bought a collared t-shirt the system might recommend more collared t-shirts in different colours for you.
A collaborative approach on the other hand compares the user’s choices with those of other users. For example if user A likes bread and user B likes bread and cheese the engine will suggest cheese to user A.
Companies choose their preferred approach based on their needs. They may also use a hybrid approach that combines the above or use other approaches as well.
The Role of UX in recommender systems:
Converting Customers: When someone is unsure about what they want to do, all they need is a little nudge in the right direction. The UX of recommender systems provides just that. It helps customers take an action, be it buying something or signing up for a newsletter/subscription. This essentially converts a window shopper into yet another customer for your business.
Increasing Sales: Providing suggestions that a user can relate to makes them feel comfortable. If you like fine-dining but all the recommendations you get relate to junk food you may move off the platform. Users need to feel at ease and UX helps with this. Keeping the user engaged with relevant recommendations allows them to spend more time on the product that they may be interested in, as opposed to spending their time searching. This increases the probability and value of a purchase.
Understanding The Customer: Recommendation engines work by learning peoples behavioral attributes. A recommender system with good UX will be able to grasp the nuances of user behavior such as societal contexts or emotional satisfaction. Better UX leads to more accurate the results leading to more orders and an increase in the number of items per order.
Driving Traffic Towards Your Website: Recommendation engines can predict user behavior. Take for example the dictionary feature on a smart phone. The phone makes suggestions based on the first few alphabets you type thus making your experience of typing a message smoother. Predictions can drive traffic to your website.
Market Insights: UX allows the engine to learn your behavior patterns. This lets companies get greater insights into their market. It’s a cyclic learning process wherein companies use this knowledge to improve their product and the engines continue to learn from and understand the customer. UX allows companies to take decisions on the pricing of their product or the features they should add or subtract and even new markets they can target.
Targeted Results: UX helps the engine decide what to recommend and how many suggestions to recommend. If for example it is observed that out of ten recommendations users usually clicked on the first three, the company may decide to show their users only four recommendations at a time, leading to more targeted results. Similarly, what content to recommend depends on how the user behaves. If the user has just bought a car s/he may not buy another one soon. So it doesn’t make sense to show a recommendation list full of cars. However, there may be an accessory for the car that the user would be interested in.
Recommendation engines are extremely powerful in today’s world. They can define the failure or success of a product or business. They also define social trends and peoples choices. You may not be a fan of classical music, but if it’s recommended to you, you might just try it out.
These systems are difficult to implement but save companies huge amounts of money and build customer loyalty. Almost all brands in the future will need to have some sort of recommendation system because at the end of the day, the recommendation engine has replaced the shopkeeper who knew you and your entire family.