In 2017, I picked up a satirical science fiction novel book called Qualityland that described a shopping situation we might be hurtling towards as we increasingly involve AI into our ecommerce practice. In the book, TheShop is the dominant online marketplace, akin to Amazon, where customers can buy everything they need. However, in this new world, packages are dropped at consumers’ doorsteps without shoppers needing to browse web pages or press any buttons. This is because TheShop's highly advanced AI algorithms can predict customers' needs even before they are aware of them, and products are automatically shipped to individuals based on their preferences and profiles.
A robot picture made with DALL-E
Although I personally still need to go through the archaic practices of browsing and clicking the “buy” button in 2023 for my shopping experiences, there are many shopping platforms and brands, primarily in China, that are already embracing a future of shopping, infused with AI. Senior executives at large retail and e-commerce businesses need to be aware of the transformative potential of AI and the implications for their organizations. By understanding these developments, senior executives can better prepare their businesses to thrive in an AI-driven retail landscape.
5 ways to integrate AI into your ecommerce business
Automating Content Generation and Product Descriptions
Perhaps the lowest-hanging fruit of process automation is using AI to automate away the repetitive and tedious tasks associated with product description pages. AI-driven natural language processing (NLP) technologies can automate this part of the ecommerce equation. There are now hundreds of tools that will enable a brand to automate photography. Instead of hiring a product photographer and taking different angels or shots of a product, one can use Snappr to change the background of certain products. IKEA, the world-renowned furniture retailer, has been using AI and 3D rendering technologies to create realistic images of their products for their catalogs and online store. In fact, IKEA has reported that around 75% of their product images are now computer-generated rather than photographed traditionally. This has allowed them to save resources and efficiently produce a vast number of high-quality images for their extensive product range.
Clothing brands are used to hiring models for photo shoots. You can now use AI to develop and change the models you use for diversity and size purposes. For example, Levi’s has been using AI-generated models for its photo shoot campaigns. In China, many brands have already started using virtual influencers instead of humans for celebrity endorsement. They have been able to save money and take control of how to craft a “human” voice to their brands.
Virtual influencer Angie, known as "Ah Xi" in China advertises for an ice cream brand
Before, brands needed to hire copywriters to develop catchy and enticing copy that describes the product. By using simple tools such as ChatGPT and Jasper, brands can now outsource this type of work or drastically increase the productivity of the current copywriter employee. We will see a huge increase in the speed of product launches, something that is more and more needed in the ecommerce space today.
Content management is also a pivotal part of an ecommerce website as the copy raises the SEO of an ecommerce site. By analyzing existing content and understanding its structure, AI-powered content generation tools can create coherent and engaging text that aligns with a brand's messaging and tone. For example, Writesonic is an excellent tool for generating blogs for ecommerce stores. For e-commerce businesses, this can result in more consistent and engaging product descriptions, improving the shopping experience for customers and potentially leading to increased sales. Additionally, automating content generation can save time and resources that would otherwise be spent on manual content creation.
Enhancing Personalization and Customer Experience
When I open my Amazon app, it will look entirely different than your Amazon app. This is because the platform has recorded your user behavior and strives to provide you with a personalized experience, showing you products that you have already seen before to increase the chances of purchase. Alibaba, an ecommerce giant in China, does something similar.
Multiple parameters that train Alibaba’s machine learning algorithm
Alibaba uses machine learning to categorize its users based on several dimensions. This chart explains some parameters that the platform uses to define its users. When a new user logs in and adds a phone number, the app already knows some basic information of the user: where they are located and maybe some pieces of data such as income from integrating with Alipay. Next, every click the user provides, “save product”, favorite, add-to-cart, purchase, repurchase or return are all actions recorded by the platform. Based on these pieces of information, the app will then reserve certain products that the user has saved or liked as re-exposure. Personalization in Alibaba’s world is also called in Chinese 千人千面 qiān rén qián miàn which means “thousands of faces for thousands of people”. The “face” means how the app is displayed.
In 2019, personalization 2.0 was enacted called 万人万面 wàn rén wàn miàn,or “ten thousand faces for ten thousand people”. This basically adds in other factors such as place and time. For instance, when you are at work, and you are looking at the P&G store, perhaps what is displayed are some office snacks. When you are at home, perhaps what is displayed are some shampoo products. The platform recognizes that its users have different needs at different times of day and different locations. The subtle differentiation ultimately leads to more fine-tuned display of products adding to the personalization catered to the user.
Flagship 2.0, a function that was introduced in the last couple of years means to change the look of a website from day to night or depending on location of the person. For instance, when you are looking at a beauty brand in the morning, maybe they will display foundation, because there is a greater need for the user to use foundation in the morning. At night, perhaps, the website will display face masks, an item that people use before they go to bed. A store front should also be able to adjust itself depending on who is viewing the store. To go back to the P&G example, if I were looking at the store, perhaps what is displayed are their top selling woman’s products. If my brother were to look at the same store, he might be viewing their top selling men’s products.
The crux of continuously perfecting this algorithm is that everyone will always be looking at a personalized version of app, specific to this person’s needs, instead of providing all users with one standard product. This will increase sales and re-purchase rates since users will be more likely to return to an app that caters personally to them.
Dynamic Pricing
Pricing is perhaps the trickiest part of the ecommerce equation since it will fluctuate your P&L if not managed correctly. By using machine learning to create a dynamic pricing system, ecommerce businesses can maximize profits and customers. We can use dynamic pricing to quickly respond to changes in consumer demand, drive revenue, enable more accurate SKU pricing and optimize profit per product.
The price of an item will depend on several large factors including supply and demand, market trends, competition and industry standards, consumer expectations, and inventory levels. Have you ever added an item to your Amazon cart only to receive a notification later that the price has changed? A prime example of dynamic pricing is Amazon's pricing model, which continually updates product prices. The algorithm processes vast amounts of data, including market trends, competitor pricing, and consumer habits, to maximize product sales and profits.
Dynamic pricing can even apply to brick and mortar retailers. Hema is a grocery chain in China owned by Alibaba. While it has an online shopping app, the management and orientation of its physical stores is at the heart of what the company calls New Retail. Western grocery chains like Tesco have sent teams to Alibaba’s headquarters to study the way Hema manages its stores. Specifically with the concept of dynamic pricing, on every shelf, each product uses an electronic price display. The product name and product description is written on the plaque along with its price. If a certain SKU needs to have its price changed due to any of the above market factors we listed, the price shown on the plaque will change accordingly. This maximizes SKU pricing accuracy and optimizes profit per product.
Another brand that uses dynamic pricing in the eCommerce space is Best Buy. Best Buy employs a dynamic pricing strategy to adjust the prices of its products based on various factors such as supply and demand, competition, market trends, and customer behavior. By implementing dynamic pricing, Best Buy can stay competitive in the fast-paced electronics market, offering attractive deals to customers while optimizing revenue. In 2017, Best Buy reported a comparable sales growth of 5.6%, which the company attributed in part to its strategic pricing initiatives.
Instead of coding and training a viable dynamic pricing system, companies can leverage existing software providers like Priceshape to plug into their site to enable dynamic pricing. A brand might also elect to choose to create its own pricing system depending on the different parameters that are relevant to its business. No matter which way the company decides to go, incorporating this strategy via AI can have significant impact to the bottom line.
Search-Based shopping versus Discovery-based shopping
We can’t talk about AI without talking about Tik Tok, as Bytedance has created a machine learning engine that has propelled the app to the top of the social media chart. Now, in addition to knowing what its users want to see, Tik Tok is beginning to break into the ecommerce sphere by knowing what its users want to buy. However, they have a slightly different take than a traditional marketing funnel that is anchored on search-based shopping. The new concept here is discovery-based shopping.
When Douyin, the sister app of Tik Tok, emerged in China, they disrupted a type of ecommerce framework dominated by search-based shopping. A user would hop into the app when he or she already knows what to buy. Search-based shopping is guarded by high intent and a narrow degree of precision, where the main agent is the shopper.
In the Douyin world, a user might be blindly scrolling through his app, an ad might appear with a product that he might have never even heard of, but because the video was very convincing, it led to a purchase then and there. This is what Douyin had recognized, that the user today does not operate in a traditional marketing funnel. Instead, they are targeted by various stages of the purchase cycle and is prone to make the purchase during any of those phases. Discovery-based shopping is informed by a low intent and a broader degree of precision, where the main agent is the data-rich platform.
Tik Tok’s path to purchase is not linear. It’s an infinite loop, as shown below. Shopping has evolved from a simple necessity to obtain a product into a community-based discussion forum and a source of entertainment. AI can help brands identify suitable influencers on TikTok by analyzing their content, audience demographics, engagement rates, and other relevant metrics. This enables businesses to collaborate with influencers who have a genuine connection with their target audience and can effectively promote their products or services. TikTok can display ads to users who are more likely to engage with the content, resulting in higher conversion rates and a better return on investment for businesses.
Tik Tok’s Infinite Loop for Discovery-based Shopping Framework
Virtual Reality (VR) and Augmented Reality (AR) Integration
The metaverse had made huge splashes in the world of CPG in the past couple of years with fashion companies such as Nike directly purchasing metaverse agencies to create partnerships within the VR and AR space. AI can enhance the integration of virtual reality (VR) and augmented reality (AR) technologies in e-commerce, creating immersive and engaging shopping experiences for customers. By leveraging AI-driven image recognition and computer vision algorithms, businesses can create realistic, interactive virtual environments where customers can visualize products in real-world settings.
This phenomenon has been bubbling up in China for quite some time, in which shoppers are used to seeing virtual try-on mirrors at make up counters and clothing stores to enable a phygital experience for the shopper. For example, brands would enable shoppers to try on certain shades of lipstick right in their phones to increase conversion. The more interactive the experience, the more memorable the brand is for the consumer. The longer the consumer has interacted with the brand, the more likely the consumer will convert.
Magic Mirror Virtual Try On at a clothing store in China
New York startup Obsess has already created dozens of virtual store experiences for its clients that range from big name brands such as Coach, Crocs, and Alo. Furniture retailers like IKEA can allow customers to virtually place items in their homes using AR technology, helping them make more informed purchasing decisions.
By augmenting a customer’s interaction with a brand in an ecommerce setting, companies are able to create a more intimate relation with its customer. In doing so, they are also able to reach a younger consumer segment, Gen Z and Gen Alpha that are more experimental and exigent with how progressive and technology-forward a brand can be.
Utilizing AI for Ethical and Sustainable E-Commerce Practices
While AI can sometimes be cast with an evil light within technology, it is also possible to use it for social good. As the demand for ethical and sustainable business practices continues to grow, AI can help businesses identify opportunities to improve their environmental and social impact.
For instance, AI can assist in reducing the carbon footprint of logistics operations by optimizing delivery routes and improving warehouse efficiency. Moreover, AI-driven tools can help businesses monitor their supply chains for ethical compliance, ensuring that suppliers adhere to environmental and labor standards.
An example of utilizing AI for ethical and sustainable e-commerce practices is the use of machine learning algorithms by the online fashion retailer, ASOS. ASOS has implemented an AI-driven tool called "Fit Assistant" that helps customers find the right size for clothing items. This tool uses machine learning algorithms to analyze customer data, such as previous purchases and returns, and provides personalized size recommendations for each customer. By offering accurate size guidance, ASOS aims to reduce the number of returns and the associated carbon emissions from shipping, ultimately decreasing the company's environmental footprint.
Moreover, ASOS actively monitors its supply chain for ethical compliance, working closely with suppliers to ensure they adhere to environmental and labor standards. The company also leverages AI and data analysis to assess the performance of its suppliers and identify areas for improvement, promoting sustainability and ethical practices throughout its supply chain.
The Future of AI in Ecommerce
Since the advent of ChatGPT in Nov 2022, the world has exploded with AI tools. In the future, ecommerce business owners can employ autoGPT to outsource a plethora of ecommerce workloads that were otherwise once performed by humans. From content production to pricing models, to VR/AR technologies, AI can optimize the ecommerce journey every step of the way. At the core of it all, our overarching goal is to improve the experience of the customer so that they become loyal fans and ambassadors to the business. We might see the day where TheShop will drop packages at our doorstep, but until then, our role as consultants and practitioners is to better educate ourselves on what AI can do to advance ecommerce.
コメント