Beauty brand collective drives conversion with product quizzes & personalized recommendations

A leading beauty company operating a collective of three premium brands was looking to increase engagement and drive higher conversion rates across its brand websites. By leveraging strategically designed product quizzes, the company was able to collect customer behavior and preference data. Their team was then able to use the Cordial platform to send various recommendation campaigns with high levels of personalization. Such strategies included “Mega” Message  Algorithms and Dynamic Product Recommendations.

Product Quizzes

These quizzes are strategically designed to captivate visitors by guiding them through a series of questions aimed at identifying the right products for their needs and preferences. By seamlessly integrating these quizzes into their online experience, the company not only enhanced user engagement but also collected valuable customer data.

Product quiz results across two of the three brands:

• 1.97M total impressions
• 3x average conversion rate

Now equipped with valuable customer data from product quiz results, the marketing team established a direct line of communication with their audience, allowing for targeted marketing campaigns and personalized product recommendations. These strategies ultimately drove higher conversion rates and customer satisfaction.

The “Mega” Message Algorithm

This block was designed so that the marketing team has the ability to create highly personalized content in their email campaigns, and the ability to control the recommendation strategy they want from the Sculpt UI using simple drop-down editors.

Several algorithms were made available in the block, including:

• Recently Carted Items (Dynamic)
• Viewed Products (Dynamic)
• Category Recommendations (Dynamic)
• Bestsellers (Fixed)
• New Arrivals (Fixed)
• Static (Fixed)

The prioritization of these algorithms can be shifted around on the fly, allowing the team to quickly create personalization waterfalls that serve personalized content to customers. The team has the ability to select two personalized algorithms and a third fallback algorithm based on fixed product recommendations.

Category recommendations inform which products should be recommended to a customer, based on multiple data points. Cart and browse events are sifted through on a product level, which allows the company to generate a score at a category level for each customer. This score is dependent on a few factors:

• The length of time since a product in that category was viewed, which declines over time. A customer viewing a product 60 days ago is a weaker signal than viewing one 7 days ago.
• The ‘strength’ of the event. A customer abandoning their cart is a stronger signal of intent than simply viewing a product, so this action is a larger contributor to that category’s score.

The algorithm sifts through the product feed and ranks each product by category score for each customer, as well as sales performance of the product over the last 30 days. Customers are then served up a set of recommendations based on their strongest category signal, and the strongest performing products within that category.

To prevent any situation where customers see too many similar products, the “mega” message block also monitors and limits how much of one category can appear at one time, allowing the brand to recommend a variety of suitable products to each customer.