In the first episode, we saw how Ocado tailors the display for each customer starting with their very first visit. But personalizing the display isn’t enough: how can we ensure that each individual receives recommendations that are truly relevant to them, rather than just those for a specific segment? That is the key challenge of applying AI to repeat purchases.
Customer segmentation alone is not enough to offer a hyper-personalised and relevant experience. When you conduct an RFM (recency, frequency, monetary) segmentation, how can you be sure that everyone in your “star” cluster is interested in the product or offer that you have chosen to present to the group? What if the customer does not want the product, or worse, doesn’t like it?
This point is key because 80% of customers want personalized experiences.[1] However, the study also suggests that personalisation is a core attribute: customers take it for granted, but if a retailer gets it wrong they may turn to the competition.
It is impossible to manage every individual situation “by hand” for each of your visitors. It is also very difficult to spot and analyse all the subtle signals that enable you to get to know each customer.
Artificial Intelligence technologies dedicated to e-commerce are capable of detecting the purchasing intention of each client, according to their behaviour and buying habits on your website. This is an opportunity to offer each customer the right products, promotions or ranges. AI learns all about your customers’ behaviours to improve the relevance of recommendations, and ultimately your conversion rates.
For Ocado, the use of these technologies has enabled the company to personalise product recommendations and send tailored messages to each customer. By doing so, Ocado has increased conversions by eight times and the most valuable shopping baskets by 20%.[2]
In the context of recurrent purchases and food shopping, product recommendations alone are not enough, as they cover only a small percentage of the shopping experience. The value of hyper-personalised shopping for the end customer and for the retailer relies on one important point: streamlining the shopping experience without allowing the consumer to be restricted to their shopping list.
The ‘shopping list’ effect means that customers confine themselves to buying only their usual products. They are not able to see the full range of what’s on offer to them, nor the “new in” products. For the retailer, this approach prevents any impulse purchases and limits opportunities for increasing basket size.
The hyper-personalisation of the shopping experience must allow for quick suggestions of products which correspond to the wants and needs of each customer. This means moving away from existing purchases and proposing promotions and products that customers did not know about or had not planned to buy, but which correspond to their customer profile.
66% of customers expect companies to understand their needs and expectations.[3] Triggering an impulse purchase means striking a balance between what the client wants to buy and what they might be interested in finding out about, while taking into account the brand’s sales strategy.
The search engine is the best place to communicate with the customer in a relevant and sales-focused way. It is the only part of an e-commerce site where the user expresses what they want and it is used a great deal in repeat shopping. In the grocery sector, 80% of revenue comes from search, which has become the primary place of purchase for customers.
Given highly relevant results, the customer will always find the product that they want. AI, as well as optimising the accuracy of results, provides an enriched search experience by suggesting intelligent and individualised search terms, categories, offers and products. The goal remains the same: to make the shopping experience ultra-efficient by ensuring that your product range is discovered and that impulse buying is encouraged.
Our example this time is our client Placedumarché.com, which specialises in the home delivery of food products. The search engine is built on AI algorithms and presents individualised search results for each customer according to their behaviour. So, in our example below, the results are well matched to particular purchasing preferences (such as products from organic farming) or special dietary requirements (halal, vegetarian, etc).
We’ve explored the full potential of hyper-personalization in the context of repeat and grocery purchases. But what happens when the customer isn’t quite sure what they want to buy? That’s the fascinating challenge of impulse buying, and our next episode is entirely dedicated to it.
[1] A McKinsey study conducted in the United States among more than 1,000 consumers
[2] Chiffres Dynamic Yield
[3] Source Salesforce
Schema: Tinyclues
Photo credit : ©Shutterstock