22 January 2024
Pokupon Case: +49% in Website Traffic and +16% in Revenue Thanks to Personalized Recommendations
About the Project
Every fan of discounts in Ukraine is familiar with Pokupon and SuperDeal, large marketplaces for discount offers. The services offer lucrative coupons and certificates that can be used to save on services in popular industries:
- travel and recreation;
- health and sports;
- self-development, etc.
You can also find promo codes for purchases in various stores here.
The total number of registered users of the two platforms reached 4 million in 2022.
'Our customers are 62% women and 38% men. These are people with an active lifestyle who want to get a good service at an affordable price. Our audience is quite diverse: working people, mothers on maternity leave, and students. But all of them have one thing in common: they want to save money.'
Valeriia Kalashnikova, Head of Retention
Service users can follow current discounts in several channels:
- on websites;
- in the mobile application;
- in Viber, push, and email campaigns.
'The task we were solving in the email channel was to automate routine processes and improve customer and brand interaction.'
Valeriia Kalashnikova, Head of Retention
To prevent customers from getting lost in the sea of offers and to help them find suitable options faster, we needed to personalize communication with the audience in email campaigns. The team of Pokupon and SuperDeal worked together with Yespo specialists to achieve this and formed two important goals:
- Increase audience loyalty with relevant offers.
- Increase sales through targeting.
Previously, the company had several types of promo campaigns:
- daily morning email campaign with top-selling discounts;
- thematic email campaign: one day of the week – one category (tourism, food, etc.);
- email campaign with limited offers (for example, for Valentine's Day, Independence Day, etc.).
The content for the emails was collected manually by marketers on the principle of "sending all the top stuff to everyone." The easiest solution for personalization was to segment the database by city – each user provided this data during registration. Many of the offers are geographically specific: some of the salons and restaurants operate only in certain cities.
However, for a communication strategy to be effective, you need to take into account a person's needs, interests, financial capabilities, etc. In addition, this data may change over time, and you need to be able to adapt to it.
It is impossible to process this amount of information and create individual offers based on the customer's history manually. Artificial intelligence can help here, analyzing customer data and behavior and generating personalized content using a predictive model.
Yespo uses algorithms that allow to generate customized offers not only in email campaigns but also on the website. Together with the customer, we decided:
- set up automatic personalization of offers in the morning and category email campaigns;
- implement recommendation blocks on websites.
Morning Personalized Email Campaign
To create emails with customized offers, we have prepared universal templates with recommendation blocks – separately for Pokupon and SuperDeal. Visually, the messages are designed in the style of each company, and the logic of content tailoring is the same.
In Kyiv, Dnipro, Odesa, Kharkiv, Mykolaiv, Zaporizhzhia, and Lviv, morning email campaigns are sent to contacts daily. This frequency can be called high, but new offers appear every day. In addition, all discounts have a lifespan on the website. Therefore, the sooner Pokupon and SuperDeal tell about the new product, the more chances the user will have to order it.
Contacts from other cities, where there are far fewer offers, receive the email campaign once a week.
Offers are generated for different user segments:
- Active – after any purchase within 14 days. Their history is the most up-to-date, and these contacts receive email campaigns with the most relevant personalized recommendations.
- Neutral – the one who performed any action (read the email, added the discount to favorites, spun the wheel or fortune in the app) within 15-45 days. This segment receives personalized email campaigns + bestsellers by city of registration (if the history is not enough to fill the email with 19 discounts). After making a purchase, such a user goes to "Active".
- Inactive – no actions have been taken for 45+ days. For these users, it is more difficult to select relevant offers because the range of their potential interests is not defined. Subscribers are sent bestsellers based on their city of registration.
Users are segmented by activity by tracking their behavior on websites through web tracking.
Web tracking technology allows you to record:
- page visits;
- adding to cart and favorites;
- price range and category of goods of interest;
- location of the website visitor, etc.
Based on these parameters, you can automatically send customized offers at the most appropriate time.
The content of the letter
Personalization starts with the message header. The email always contains dynamic content, such as the current date and city, which is set individually for each customer. This way, the subscriber can immediately see where and when the offers are available.
The main content of the message consists of 19 items with discount offers. They are divided as follows:
Main Offer, which can be a new product or a top-selling offer. For each user, this will be the right discount for them. For example, Helen likes to go to beauty salons. It will prioritize either the offer of a new salon or the offer of a salon from the bestseller category if the new discounts for that day do not have a corresponding service. The Main Offer occupies one position. The block with the discount is positioned higher than the others over the entire width of the sheet.
The following 18 offers are placed in double blocks:
- New – if the date of the offer launch is the current day, it is considered new. This information is contained in the corresponding feed attribute. New items are listed in the email after the Main Offer.
- Top discounts. Bestsellers take the remaining positions and should not be repeated with the Main Offer.
If an affiliate has several new discounts, only one will be used in the email campaign. This way, subscribers receive a variety of content, and healthy competition is created.
Dynamic content is inserted using SRT blocks.
Blocks with discounts from the same category (tourism, entertainment, etc.) are not placed in a row to diversify the offer.
In addition to the order, there are several other rules for displaying content. The email campaign takes into account the customer's purchases: the purchased discount item is no longer included in the email. Also, the recommendations do not include discount items that expire on the day the email campaign is sent or on the first and second days after it is sent.
If less than 10 recommendations are generated for a user, the email is not sent. This condition can be changed in the SRT unit settings.
After the discount offers, there is a block with popular categories.
This way, a subscriber can get to the desired section on the website with one click.
The subject line is automatically generated based on the first 3 offers in the email using the formula:
Short name of the discount + coupon price ☀️ Short name of the discount + coupon price ☀️ Short name of the discount + coupon price
The subject line is customized for each subscriber. In the inbox, the user sees an interesting discount at an attractive price:
Category Email Campaign
This email campaign is divided into 5 categories:
- food and restaurants;
The emails are sent according to a schedule: one day of the week corresponds to a certain selection of offers. For each category, we have prepared templates for Pokupon and SuperDeal.
The selection of offers from the travel section covers all 44 cities, as travel discounts are launched all over Ukraine.
Letters from the Beauty, Food and Restaurants, Health, and Entertainment categories are sent to users from Kyiv, Dnipro, Odesa, Kharkiv, Mykolaiv, Zaporizhzhia, and Lviv. Email campaigns are launched for contacts who have made at least one purchase in a given category in a certain city. Users are segmented by selecting the appropriate parameters in conditional groups.
The content of the letter
The email of the category campaign is generated similarly to the morning one. The first to be placed is the discount item that the algorithm has determined to be the most suitable for a particular user. The following positions are also arranged in two per block.
This email campaign does not take into account the newness of the discount – only the tops in the category participate in it. The algorithm decides what will be a priority for the user if they have a history of interaction. If there is no story, the customer will simply receive bestsellers.
Personalized offers for each customer
Users of the Pokupon and SuperDeal marketplaces have access to more than 8000 different discounts. To help each customer with the choice on the website, we implemented recommendation blocks. Yespo's experience shows that the ROI of product recommendations on a website can reach 1734%.
Several types of recommendations are involved in content creation:
- Personal – based on the contact's data. The block is built taking into account views, purchases, and other factors of user behavior.
- Bought together – based on product data. The algorithm selects offers based on product data: which items are chosen by people who bought this discount item.
- TOP on the website – based on the general data of the website. The block is built taking into account sales, views, adds to favorites, etc., of all users and displays the most popular items. The city must be taken into account.
- TOP in the category – based on the overall category data. The block is built taking into account data on all sales, views, adds to favorites, etc., in a particular category of discounts. Offers are selected from the highest-level category.
The page of the current discount
This page contains the following blocks of recommendations:
1. "You might be interested in" or "TOP in the category" if there is no user data.
The neural network predicts the most likely click while applying the following filters:
- from the same category;
- price range;
- from the same city.
The block consists of 2 rows of three cards in each:
2. "Bought together".
Since you can only add one discount to the cart and, accordingly, buy one, the algorithm analyzes subsequent customer purchases. For example, a user has purchased a discount coupon for tests in a particular laboratory. The neural network checks who bought this discount item in the last 7 days, which discount items were bought after it, and gives a relevant result.
Terms of display:
- offers from the same city;
- if other users have not bought anything within 7 days after purchasing the item and there is no information for the block, it is not displayed.
The expired page of the campaign
First, we recommend active discounts of this partner. If there are none, similar discounts from other partners are available.
The successful payment page
Discount recommendations are based on the customer's history and city-specific data.
The "Favorites" page
If the user does not have any discount items added to the favorites list, recommendations with the top three discount items on the website are displayed for them. If there is already at least one item in the favorites, the block is not displayed.
A page with a small number of discounts
The block is displayed on subcategory pages with a small number of discounts. For example, in a certain city, there are few offers in the "Pies" section:
- if there are ≤ 3 discounts on the page, we recommend items from the parent category (Food Delivery);
- if there are 3 to 6 discounts in a category, we recommend a category above the parent (Food and Restaurants).
Among personalized email campaigns, morning messages are the most effective. They bring 55 to 90% of sales generated by emails with individualized recommendations.
According to the results of testing the first three months of the email campaign, Pokupon managed to achieve the following growth rates:
- website traffic: +49.1%;
- active customers: 17.7%;
- number of transactions: +12.1%;
- revenue: 15.6%.
After receiving good results from Pokupon's test email campaign, personalized communication was launched for SuperDeal subscribers.
The recommendation blocks on the pages of both marketplaces began to bring in 2-5% of all sales made on the websites.
Plans for the Future
The Yespo team continues to work on recommendations for SuperDeal and Pokupon. Customized offers will appear on other pages to make the user's journey even more comfortable. The company also plans to launch communication with the audience through App Inbox.
If you think that personalized communication is exactly what your business needs right now, leave a request for a free consultation. We'll help you choose the right Yespo platform tools to help you achieve great results.