Lotte and Docomo Succeed in Targeting PoC for Coupon Distribution Using Docomo's Virtual Marketing Technology
Lotte and NTT Docomo successfully conducted a proof-of-concept (PoC) for targeted coupon distribution. By utilizing an LLM to generate 'virtual customers' representing non-buyers of Ghana chocolate, they achieved up to a 1.76x higher purchase rate compared to general targeting, validating the effectiveness of virtual marketing technology.
📋 Article Processing Timeline
- 📰 Published: April 22, 2026 at 19:00
- 🔍 Collected: April 23, 2026 at 00:02 (5h 2m after Published)
- 🤖 AI Analyzed: April 23, 2026 at 01:16 (1h 14m after Collected)
Lotte Co., Ltd. (hereafter Lotte) and NTT Docomo, Inc. (hereafter Docomo) announced today the successful results of a proof-of-concept (hereafter PoC) for targeting in coupon distribution. The PoC utilized Docomo's newly developed virtual marketing technology (*2), which generates virtual consumer models (*1) (hereafter virtual customers) using large language models based on various corporate data and allows for interviews with these virtual customers. This was combined with data from Docomo's approximately 100 million (*3) d-account members and their purchasing data. Notably, this is the first time a PoC for targeting in advertisement and coupon distribution has been conducted using virtual marketing technology.
In this PoC, coupons for the "Ghana Milk Chocolate Series (*4)" were distributed to two groups: customers who have attributes similar to virtual customers generated by the virtual marketing technology under the premise that they have never purchased the "Ghana Milk Chocolate Series," and randomly selected general customers. The purchase rates of the "Ghana Milk Chocolate Series" among the customers who received the coupons during the PoC period were then compared. The results showed that the purchase rate among customers similar to the virtual customers generated by the technology was up to approximately 1.76 times higher than that of general customers. This confirmed that utilizing virtual marketing technology enables an understanding of customers who have never purchased the "Ghana Milk Chocolate Series" and allows for effective approaches based on that understanding.
[Overview Diagram of the PoC]
1. Background
For companies, understanding consumers is a crucial process in market research for marketing. Currently, many companies understand consumers through product shipment volumes and customer surveys, which poses a challenge in executing effective targeting. Particularly when approaching customers who have never purchased a specific product (hereafter non-buyers), there is a lack of information such as purchase history. Therefore, Lotte felt a challenge regarding the low level of understanding and targeting accuracy for non-buyers.
To address this challenge, Lotte and Docomo believed that generating and interviewing virtual customers on the premise of being non-buyers using "virtual marketing technology," and extracting customers with similar attributes from approximately 100 million d-account members, would enable a more sophisticated understanding and targeting of non-buyers. Consequently, this PoC was conducted to verify whether understanding non-buyers and taking effective approaches based on that understanding could be made possible by utilizing virtual marketing technology.
2. Overview of the PoC
This PoC was conducted from Thursday, January 15, 2026, to Saturday, February 14, 2026. Initially, based on data linked to d-account members such as gender and age, as well as purchase data held by Docomo, 1,240 virtual customers who have the premise of never having purchased the "Ghana Milk Chocolate Series" were generated using the virtual marketing technology. Subsequently, by interviewing the virtual customers regarding chocolate purchasing behavior, such as their frequency of buying chocolate and experience in making sweets with chocolate, three types of clusters (*5) were created: "Price-Oriented," "Preference-Oriented," and "Awareness-Oriented." A total of approximately 2 million customers with attributes similar to each cluster were then extracted.
Coupons were distributed to these customers and randomly selected general customers within the "d-barai®" app and the d-point club app. By comparing the rate of users who clicked the coupon banner to display the coupon (hereafter coupon display rate) and the rate of users who actually purchased the product (hereafter purchase rate), a verification was conducted to see if virtual marketing technology enables the understanding of customers who have never purchased the "Ghana Milk Chocolate Series" and facilitates effective approaches based on that understanding.
Additionally, before actually distributing the coupons for the three clusters created this time, predictions of the "Expected Response Rate"—which indicates the probability of customers taking action regarding the coupon, such as viewing the banner, displaying the coupon, and purchasing the product—were made for each cluster using the virtual marketing technology. As a result, the hypothesis was established that the "Expected Response Rate" for the "Preference-Oriented" cluster, which has the characteristic of emphasizing personal taste, was low, and that the coupon display rate and purchase rate of customers similar to the "Preference-Oriented" cluster might also be low in the actual distribution.
[Expected Response Rate of Clusters]
3. Results of the PoC
As a result of this PoC, in the "Price-Oriented" cluster, which emphasizes price, the response to the coupon was the highest, with a coupon display rate 1.66 times higher and a purchase rate 1.76 times higher compared to general customers.
In this PoC, coupons for the "Ghana Milk Chocolate Series (*4)" were distributed to two groups: customers who have attributes similar to virtual customers generated by the virtual marketing technology under the premise that they have never purchased the "Ghana Milk Chocolate Series," and randomly selected general customers. The purchase rates of the "Ghana Milk Chocolate Series" among the customers who received the coupons during the PoC period were then compared. The results showed that the purchase rate among customers similar to the virtual customers generated by the technology was up to approximately 1.76 times higher than that of general customers. This confirmed that utilizing virtual marketing technology enables an understanding of customers who have never purchased the "Ghana Milk Chocolate Series" and allows for effective approaches based on that understanding.
[Overview Diagram of the PoC]
1. Background
For companies, understanding consumers is a crucial process in market research for marketing. Currently, many companies understand consumers through product shipment volumes and customer surveys, which poses a challenge in executing effective targeting. Particularly when approaching customers who have never purchased a specific product (hereafter non-buyers), there is a lack of information such as purchase history. Therefore, Lotte felt a challenge regarding the low level of understanding and targeting accuracy for non-buyers.
To address this challenge, Lotte and Docomo believed that generating and interviewing virtual customers on the premise of being non-buyers using "virtual marketing technology," and extracting customers with similar attributes from approximately 100 million d-account members, would enable a more sophisticated understanding and targeting of non-buyers. Consequently, this PoC was conducted to verify whether understanding non-buyers and taking effective approaches based on that understanding could be made possible by utilizing virtual marketing technology.
2. Overview of the PoC
This PoC was conducted from Thursday, January 15, 2026, to Saturday, February 14, 2026. Initially, based on data linked to d-account members such as gender and age, as well as purchase data held by Docomo, 1,240 virtual customers who have the premise of never having purchased the "Ghana Milk Chocolate Series" were generated using the virtual marketing technology. Subsequently, by interviewing the virtual customers regarding chocolate purchasing behavior, such as their frequency of buying chocolate and experience in making sweets with chocolate, three types of clusters (*5) were created: "Price-Oriented," "Preference-Oriented," and "Awareness-Oriented." A total of approximately 2 million customers with attributes similar to each cluster were then extracted.
Coupons were distributed to these customers and randomly selected general customers within the "d-barai®" app and the d-point club app. By comparing the rate of users who clicked the coupon banner to display the coupon (hereafter coupon display rate) and the rate of users who actually purchased the product (hereafter purchase rate), a verification was conducted to see if virtual marketing technology enables the understanding of customers who have never purchased the "Ghana Milk Chocolate Series" and facilitates effective approaches based on that understanding.
Additionally, before actually distributing the coupons for the three clusters created this time, predictions of the "Expected Response Rate"—which indicates the probability of customers taking action regarding the coupon, such as viewing the banner, displaying the coupon, and purchasing the product—were made for each cluster using the virtual marketing technology. As a result, the hypothesis was established that the "Expected Response Rate" for the "Preference-Oriented" cluster, which has the characteristic of emphasizing personal taste, was low, and that the coupon display rate and purchase rate of customers similar to the "Preference-Oriented" cluster might also be low in the actual distribution.
[Expected Response Rate of Clusters]
3. Results of the PoC
As a result of this PoC, in the "Price-Oriented" cluster, which emphasizes price, the response to the coupon was the highest, with a coupon display rate 1.66 times higher and a purchase rate 1.76 times higher compared to general customers.