In 2020, the Polish branch of Carrefour began a test implementation of an email marketing automation tool. Crucial goals for the development of the company's e-commerce channel are:
A. Building customer loyalty
B. Churn reduction
The company decided to choose User.com. One of the important factors that prompted the Carrefour team to cooperate was close assistance in technical implementation and substantive support in the design and creation of specific scenarios and automation.
After identifying the customer's needs and preliminary analysis of traffic and user behavior in the Carrefour.pl online store, it has been proposed to conduct the Proof-Of-Concept (POC) process in two stages:
Adding the User.com script on the Carrefour website, identifying "low-hanging fruits", proposing specific solutions based on practical experience at working with e-commerce customers, and efficient implementation of the first automation.
After several days of data monitoring, we identified areas for optimization. Among the visits to the website, there was a relatively high percentage of users who visited the store but showed no interest and did not leave an e-mail address. Additionally, few of those who created the account made their first purchase. On the other hand, some of those who shopped for the first time never returned to the store again.
In the first step, the project team defined the data structure and, based on it, configured the necessary events. At this stage, an important task was also to import user databases from various sources to the User.com application (including the tool used by Carrefour to send newsletters) and integration with the e-commerce system. By sending information between systems via API, it was possible, among others, to update information on consent to marketing communication in real-time. After the initial technical integration, the User.com team recommended the implementation of the following automations:
- Welcome email campaign (sequence of several messages) after creating an account,
- Abandoned Process Rescue Campaigns: Abandoned Page and Abandoned Cart.
Automation optimization from stage 1, deepening technical integration and implementation of further automation paths.
At this stage, the team focused on improving the effectiveness of the solutions implemented earlier - a number of A / B tests were carried out, analytics were refined and event reporting to Google Analytics. An important step was also importing historical data about users who have been inactive for a long time to the User.com application. Thanks to this, Carrefour and User.com teams could implement comprehensive anti-churn activities, including reactivation of former clients. The API integration between the User.com application and the e-commerce platform was deepened - thanks to this, the automation and analytics included not only front-end events but also those sent from the backend. To be able to direct activation and loyalty activities, the effectiveness of which was confirmed in stage 1, to the growing number of users, an important element of stage two was the implementation of automation, which will help to repeatedly increase the base for sending e-mail campaigns.
It is worth pointing that from the very beginning, design teams were established on both sides, and the main assumption was close cooperation between the teams at every stage. We agreed that for joint work to bring tangible business benefits to both parties, the basis should be a combination of competencies of both teams, efficient and flexible operation, and honesty.
Thanks to mutual efforts, already at the beginning of the first stage of POC in Google Analytics satisfactory revenues generated because of the implemented automation were documented.
Having in mind the crucial goals of Carrefour e-commerce, the team jointly proposed specific solutions.
Collecting detailed data on user behavior
To be able to direct personalized messages to the user at the right moment, a very important starting point was the development of the data structure. The project team prepared a list of events and data that had to be configured to monitor user actions (such as clicking a given CTA button, opening an email or clicking a link, or interacting with a chatterbot) and subsequent evaluation of the effectiveness of individual automation
At the beginning, a group of events related to website behavior (including standard purchasing behavior, such as view, add_to_cart, purchase) was implemented, and along with the implementation of subsequent automation paths - tags were added to track important elements of each path. The embedding of the tags was carried out using the Google Tag Manager, in turn, to analyze the effectiveness of individual activities, dashboards on User.com and Google Analytics with a configured attribution model were selected
Saving Abandoned Page and Abandoned Shopping Cart
To respond to the need to activate customers and reduce the number of people who give up shopping in the store, the team recommended the implementation of automation-saving abandoned processes. These automation pathways were implemented early in the first stage of the POC in response to the identified "low hanging fruits". The team wanted to verify the effectiveness of the recommended paths as soon as possible and confirm that these activities would bring additional income.
During the initial research and analysis of the main problems faced by the Carrefour e-commerce team, it turned out, among others, that many people visiting the store go to several subpages, browse products, and then leaves the website. The solution proposed by the design team was to send an e-mail thanking users who viewed product subpages for the visit but left the store without adding the product to the cart. The campaign aimed to encourage the user to return to the store and make a purchase. To limit the number of messages sent to a single user, capping has been introduced (more about managing communication directed to a single recipient and the applied capping strategy can be found in the section "One-time promotional campaigns and the importance of detailed segmentation").
The second path was to save the interrupted purchasing process. When the user, after adding products to the cart, left the store's website without finalizing the transaction, he was added to the list of users to whom the e-mail campaign with a reminder and an incentive to return to the store as directed. A great advantage of email creations in this campaign was personalization - the messages displayed products that the customer had left in the basket.
The essential part of creating automation saving abandoned processes was taking into account the customer's perspective and ensuring the best Customer Experience. Both automation paths worked in synergy - for example if the user on that day was sent an abandoned cart rescue campaign, they no longer received a message aimed at rescuing the abandoned page.
Newsletter subscription with special discount coupon
Just at the first stage of POC, it turned out that the rescue abandoned processes campaign brought good results. The obvious recommendation was to propose activities aimed at increasing the base of users who left their email addresses and consent for communication. To increase conversion and encourage new users visiting Carrefour's website to leave their e-mail, and the next step- to set up an account, the team prepared a pop-up with a newsletter subscription, which was intuitional from the user’s view and at the same time based on segmentation.
To be compliant with GDPR standards, the final version of pop-ups contained the “double opt-in” element- a two-stage process was created with the clearly marked moment of consenting to marketing communication. As an incentive a discount on a first purchase was used.
Pop-up with newsletter subscription - 1. step
Pop-up with newsletter subscription - 2 step
It was really important to display pop-up only to a specific group of users and to prevent giving coupons to third parties.
Thanks to the accurate tracking of users' behavior and collecting the history of their interactions with the website, the team prepared an accurate segmentation of users based on attributes and events - the pop-up was set to be displayed only to users who are not yet subscribers to the newsletter.
To minimize the risk of abuse related to discount coupons, it was needed to integrate the e-commerce system used by Carrefour with user.com. Thanks to two-way data transmission via API, the team was able to assign each generated coupon to a user’s specific unique e-mail address ( i.g. a unique user on the side of the online store). What's more, thanks to full integration, the team was able to monitor whether a given coupon was used. Each discount coupon had a limited validity period. One of the recommended next steps was to use this knowledge to send automatic reminders about the expiry of the discount, which was to cause the FOMO (Fear Of Missing Out) effect and additionally support the conversion.
Welcome email campaign after account creation.
When signing up for an account at Carrefour, a multi-stage mailing campaign was launched (a sequence of several messages sent at specific times). In the first message, sent as part of the campaign, users received an invitation to take part in a short survey - thanks to this, the team gained additional information about customers' eating and shopping preferences, which in turn allowed for even more accurate segmentation of the user base and more precise personalization.
After the success of the welcome campaign in the first phase of implementation, the project team proposed to expand the original path and enhance it with a strong loyalty element.To reduce the risk of customers deactivating after their first purchases, a complex automation path was built where new customers received a welcome email with information about discount codes. The campaign proved to be successful - it clearly increased the percentage of customers who make second and subsequent purchases.
After a few weeks of testing the campaign and analyzing the results, it was time for optimization. The team conducted numerous A/B tests - they checked the effectiveness of different email subject lines (Open Rate optimization) and different images (Click Through Rate optimization).
Importantly, with the aforementioned integration of the e-commerce system and User. com, the team set up two-way marketing consent updates.
Further improvements were also proposed, which, when implemented, thanks to the transfer of data using API would ensure that information about consent given (or withdrawn) for marketing communications would be updated in real-time, regardless of when the user made the decision.
Anti-churn activities
In response to Carrefour's second crucial e-commerce goal, the team proposed implementing measures to rescue "dormant" customers. We carried out a mass mailing, and in addition, we also created automation for recurring mailings, with messages to inactive customers.
The reactivation effort was based on audience segmentation using two main criteria:
The time from which a customer remains purchase inactive,
Average basket value.
Consider the data mentioned above, the team divided the users into several segments and prepared different mailing creations (different content and different values of discount coupons) to maximize the chances of reactivation while taking into account the cost aspect.
Analytics
This section was particularly useful for detailed analysis of newsletter subscriber data - it monitored, among other things: growth of the newsletter base; the number of new newsletter signups coming from automation with the pop-up; what proportion of subscribers made purchases in a given period; who unsubscribes from the newsletter; whether the users who unsubscribe are mostly those who have recently joined or long-time subscribers.
Additionally, by adding events at various stages of interaction with the chatbot (more about chatbot at the bottom of the page), the team was able to analyze the attractiveness and usefulness of this channel for online shoppers.
Just a few days after implementing automation to rescue abandoned processes, Carrefour's team noticed additional revenue generation in Google Analytics. Significantly, with the optimization of campaigns, creation of further automation paths, and implementation of User.com specialists' suggestions, the positive trend of additional revenue growth not only continued but clearly accelerated. In February 2021, revenue coming directly from User.com partnership grew 47% month-over-month, in March month-over-month growth was 27%, and in April: 88%.
Chatbot Kerfuś
Carrefour has also decided to implement a chatbot. Using User.com's intuitive drag-and-drop automation creator, the client team built a very sophisticated pathway on their own, with multiple questions and pre-defined answer choices.
One of the challenges of Carrefour's customer service team since the outbreak of the COVID-19 pandemic has been the constant overloading of the helpline - consultants have been unable to handle all calls, which has harmed customer sentiment.
By developing an extensive chatbot path and covering the main issues customers face, the team was able to significantly ease the burden on the hotline and strongly improve the speed of response to customer inquiries.
The project team configured additional events that were activated when the user performed a specific action (e. g. , starting and ending a conversation with a chatbot). It allowed users to create dashboards and widgets with graphs in the Analytical section of User.com application and enabled quick monitoring of chatbot's effectiveness as well as the evaluation of the quality of help by users.
One-off promotional campaigns and the importance of detailed segmentation
The User. com platform has also proven itself as a tool for conducting one-off promotional campaigns and informing customers about temporary price reductions. Thanks to a survey conducted at the beginning, the team had information about users' shopping and eating preferences and could adapt communication and products recommended in newsletters to individual needs of customers.
Customer segmentation, based on events recorded for them (purchases, newsletter sign-ups, etc. ), was key to the activities undertaken. An example would be periodical newsletters, the content of which was conditioned by the types of purchases made by a given customer (food, non-food, mix). Segments were dynamically recalculated before sending so that the user received messages that matched their behavior and product preferences.
A beneficial solution implemented with the help of User. com consultants was multidimensional capping, the purpose of which was to dynamically limit outbound communication to a single user.
At certain times (e. g. , after the first message from the welcome campaign was sent to a given user), the system automatically added tags associated with particular campaigns to the user, which were removed after a few days. On the other hand, before sending each email campaign, the system verified whether a given user is currently assigned a specific tag, and on this basis it decided whether this campaign should be sent to this person. With this strategy, the teams eliminated the problem of sending campaigns to customers too frequently, while ensuring that key messages (like an email campaign with discount codes) were sent to the recipient.
Replenishment campaigns
Another of the activities was the implementation of a replenishment campaign. It was aimed at people who purchased products that require regular replenishment, such as razors and diapers for young children.After 30 days of purchasing the product, the system checked if the customer had restocked over the last month, and if not - the customer received an email with an incentive to buy again. To provide the best possible Customer Experience, by downloading data from the product feed, before sending a message to a given user, it was automatically checked whether a specific product is currently available in the Carrefour warehouse and can be efficiently shipped to the customer.
