Artificial intelligence for online shops
sales growth, cutting expenses, conversions increase
Thousands of online shops disappear every year only in Russia. It is easy to enter the fast growing e-commerce market but it is much harder not to be driven out of the market by constantly appearing competitors. Nevertheless, artificial intelligence technologies help online sellers solve this problem increasing their profit, decreasing the number of refusals and modeling personal experience for each customer.
It is much harder to get competitive advantages for online shops than for traditional offline outlets, which can benefit from their location close to the customers. On the internet you can find the same merchandise on dozens or even hundreds of platforms. Sales statistics show us that customers choose the most convenient e-shop. Thus, according to Shopify research, 70% of users leave the website of an e-commerce platform if they are not able to find the product they need straight away, when 12% go to a competitor if they don’t like your offer.
In the field of e-commerce it is crucial to understand each client. Of course this can not be achieved by managers, so these tasks are reassigned to artificial intelligence. Specially trained algorithms are able to indicate what customers want and what they need sometimes long before the customers realize it themselves.
Depending on the market segment, pricing and positioning each e-shop needs different customized solutions. But all of them in one way or another aim to process the data and to find a personal approach to each client.
Today we would like to share with you some stories of online shops which applied AI technologies working with their customers and got better performance.
Personalized recommendations
Every consumer searches for something on the internet, opens your e-shop tabs, scrolls down looking for some items, analyzes the prices. Every action may and has to be memorized, that does not even require any personal data.
Artificial intelligence allows you to keep track of your customers’ shopping journey, analyze which products they prefer, which requests and how many times have been made from a certain device and lets you serve your customers with personalized product recommendations.
For example, recently this solution has been implemented for “M.Video” e-shop. Managers used to make personalized offers for clients, they used to recommend their clients to buy a memory card with a smartphone or to take a bag with a laptop, but today all the offers are made by artificial intelligence.
After AI tools implementation the e-shop customers started getting personalized recommendations based on the history of their actions, as well as general rules and templates. “M.Video” takes into consideration search requests history, the ways how customers came to the e-shop website, previous purchasing experience etc. This resulted not only in cutting operators’ salaries expenses but also helped achieve 30% profit growth along with 12% sales increase.
Amazon users are no more surprised to see recommendations blocks on every page of this largest on-line hypermarket. But not many people ever think about how much money it brings to the retailer. Back to the first stage of the technologies implementation, when Amazon analysts admitted low level of recommendations efficiency (Toronto University scientists evaluation showed only 5% efficiency), Amazon sales increased by 34% and sales for one quarter after the technologies implementation reached $43,7 billions.
World famous Chinese giant Alibaba has been developing their own personalized recommendation technology — E-commerce brain for several years already. The system uses real-time data, analyzes the content seen by the customer and forms the recommendations. Based on this system the e-shops showcases which adapt to each customer are created. The first year after this technology implementation November sales conversion increased by 20%.
Smart search
Another e-commerce area where artificial intelligence technologies may be applied is search automation. Taking into consideration that internet giants such as Google and Yandex have been using artificial intelligence to filter and sort search results, internet users got used to finding what they need very quickly. That has become a serious problem for online shops with a wide range of products. Here an intellectual tool is necessary because, according to Forrester Research, 43% of users go straight to searching. And as Salesforce Demandware claims, customers who use search, spend in online shops 2,6 times more money.
For example, Martial Arts supplies store ZoovGear revealed that most of the customers don’t order anything the first time they visit the website. Sometimes customers come back 3 or 4 times and some of them up to 20 times! Also many customers preferred to call the manager as they were not able to find the item they needed on the website. And those who were disappointed with the shop just left without buying anything.
After the implementation of search — simple smart search without any additional features, sales conversion following search on the website increased by 45%. General conversion level increased by 10%. So that the new solution ensured quick ROI.
One more example of successful smart search implementation is FlipCart, a popular Indian e-shop. Buyers Application called Project Mira is used to process half of search requests, and most additional items are added by the customers to the cart in this app. The company estimates 12% sales growth as a result of AI implementation.
To win your customer back
In any case some customers leave. But many of them don’t do that because your e-shop doesn’t suit them. There are plenty of reasons for customers to leave: some people are not comfortable with the delivery terms, some might change their mind because of the colour of the item, others compare prices and go with a “100 rubles cheaper” option of another store.
There are two ways to win your customers back. Firstly, personalized offers can be made for the customers, who have left, in an attempt to “chase” them by means of programmatic ad, personal account and mailing, if the customers are registered on your website. Secondly, you can also work on the problems that the client had with the failed purchase.
“M.Video” that we mentioned earlier used both ways. By the way, that is how the company modified the delivery system adding more convenient hour time slots for delivery. Also, if an item, the customer has searched for before, appears in stock, the potential buyer immediately gets notified about it.
Productive communication
Artificial intelligence is capable of replacing an operator for basic communication with the client. When it comes to such questions as delivery terms, order status, detailed information about the product — there is no need to wait for operator’s assistance. You can integrate a chatbot to your website and use one of the social networks the same way numerous Russian companies do.
Positive effect is visible when implementing a custom-designed “digital assistant” under the condition that machine learning was applied in order the chatbot could really “understand” the customers. For example, Indian online-shop FlipCart claimed 14% increase of service recommendations and 25% decrease of customers’ dissatisfaction level after the implementation of chatbots with AI.
The price choice
“How to set up a price for the product?”, it is not an easy question to answer. How can you actually prevent clients from leaving you for your competitor because of a slightly higher price you’ve given? How not to go in the red having the lowest prices on the internet? Such online giants as Amazon use the most efficient dynamic pricing. And today AI is accessible not only for big market players.
One Russian online shop of fashionable clothes and accessories made it clear that trying to launch seasonal collections relying on managers’ advice and aspiring to catch up with the market prices manually lead to negative ROI. AI implementation helped them find the balance and adapt their pricing on a daily basis.
If before pricing changed once a month, now the algorithm is constantly suggesting new pricing options, analyzing inventory turnover, warehouse stock, product return, seasonality, day of the week. Moreover, AI allowed us to check the hypothesis about products, customers are ready to spend much more money on. As a result project margin has increased by 12% in six months thanks to the machine learning algorithms implementation.
Smart delivery
Logistics is one of the biggest problems for popular services. Under self-isolation the delivery issue turned out to be rather challenging, especially in big cities. Thus, market leaders in delivery services use machine learning and artificial intelligence technologies for demand forecasting and customer attraction.
For example, “Sbermarket’s” number of orders increased 10 times in spring 2020 and the company managed to handle the overload only using custom
designed system called Market Efficiency, which brought out the demand for couriers, based on a wide spectrum of data. As a result, the company is able to process a maximum number of orders without overpaying the couriers.
As for China, 70% of Alibaba deliveries within the country — that is 42 millions of packages daily — are supported by Cainao platform. This AI system picks up the most efficient delivery itineraries. This allows to decrease transport usage by 10% as well as to shorten delivery distances by 30%. All this lets you considerably reduce transport logistics costs.
Are you ready?
It is not difficult to implement artificial intelligence for an online shop.
Even if you are not aware of it, your company receives a huge amount of data on the website every hour, your CRM-system contains just as much useful data as well.
Here in SmartCoders we help our clients deal with their information assets, we also offer free of charge trial periods for AI models. Having estimated the benefit of algorithms for your business, you will be able to make an informed decision together with your trusted partner.
