How artificial intelligence scaling up e-retail?
E-commerce websites embrace artificial intelligence for enhanced product discovery.BY Guest author | Apr 27, 2017 | comments ( 0 ) |
Do customers often abandon their carts on your e-commerce website? Do you have plenty of instances when they search for specific styles and after spending a few minutes of scanning the options, they leave the store? If the answer to these questions is yes, then you may have a problem and no, it’s not about the inventory you stock or the styles you offer. It is a much deeper and more complex issue of product discovery.
So what does product discovery really mean? In simple terms, when a customer logs on to your e-commerce store and types ‘green cold shoulder dress’ in the search bar, she has a specific style in mind. On hitting enter, if she doesn’t see exactly what she is looking for and is greeted with generic green dresses, she is likely to look elsewhere. If you didn’t have a cold shoulder dress in store, which is very much in vogue at the moment, it’s an inventory management flaw. But it’s more disappointing and far more deleterious if despite having the right kind of styles and the correct number of SKUs, a customer is unable to discover that product. What is the solution of this mismatch, you may ask. The answer lies in machine learning and artificial intelligence.
E-commerce shopping portals need to annotate product images manually, create product titles and descriptions along with tagging them accurately. Most of the times customers search for their favourite pattern or design by typing their preferences on search engines. Therefore, in order to make a product visible, it is required of the e-commerce company to group and categorize products accordingly to help customers navigate through the store easily with their desired choice of products. Keeping in mind the search behaviour of consumers, it may be a pain staking task to use expensive human labour and effort to tag each and every product.
Here, machine learning can be of great help as it allows automated tagging. With machine learning, an e-retailer can keep increasing inventory as the items can be tagged automatically with the relevant attributes and product categories. AI uses algorithms that study the visual quality of products, automatically add multiple tags and categorize them accordingly. With AI, each of the outfits, shoes or accessories can be tagged with several relevant keywords and can also be made use of as another vector for cross-referencing of products. For example, if a customer is viewing tank tops, she could click on the tags to see a list of other products with the same tag and choose from a larger variety of options.
Machine learning, a sub segment of AI, takes data and learns by itself. One of the most promising tools in AI, machine learning has the ability to standardize product categories, quickly apply knowledge and training required from large data sets to recognize objects, speech and visuals, etc. making it easier for customers to search for their favourite products in the most convenient manner. They can, to a great extent, improve product search capabilities on e-commerce. While e-commerce search solutions initially depended on manually-fed keywords, machine learning looks beyond keywords and also considers customer ratings, click rates, conversion rates, etc. to help deliver product results that customers are more likely to buy.
In addition, product recommendations are a common practice followed by e-commerce companies today which is also an application of AI and machine learning. Based on a customer’s search patterns, the product recommendation engine identifies the preferences with respect to colour, size, pattern, rates and brands they follow to suggest similar choices. The recommendation algorithms monitor the activities done by the customer on the website and look into the past purchases to offer them similar products. That’s not it. AI offers customers a personalized experience by analyzing vast data sets much more efficiently than a human being. It is capable of analyzing a series of transactions in the shortest amount of time to drive the most differentiated customer experience.
Machine learning has enabled easy recognition of patterns and makes predictions as per the data collection. Fynd, the unique fashion e-commerce portal has achieved 87% efficiency in determining the product pattern with the help of machine learning. The firm is now working on 12 other attributes that will help to determine the attribute value of any image at an efficiency of over 90%.
With the adoption of AI and machine learning, the pace of work on e-commerce has increased, as this has led to a decrease in human dependency and intervention. Initially, customers had to choose from what is available on the platform, but machine learning has made it possible for customers to own what they want in a way that satisfies their tastes and preferences. The deployment of machine learning and AI applications have made it possible for e-commerce retailers to deliver the right kind of products at the right place and the right time. Its adoption is fast becoming indispensable and for good reason. After all, why stock a product if a customer is unable to discover it when he/she most wants to buy it?
The article has been penned down by Harsh Shah, Co-founder, Fynd
E-commerce industry to
The e-commerce market in the country will be about $ 70 billion by 2020 and players will be
- Executive Movement
Kishore Biyani resigns as
Future group chairman Kishore Biyani has stepped down from the post of Managing Director of Future
In addition to improving backend processes, retailers across the globe, including a few in India,