Optimising recommender systems for the retail industry by combining flavour and consumer insights
Retailers across the globe heavily invested in recommender systems in order to predict future purchasing likeliness based on past purchasing behaviour. Together with the Ester team, we took on an entirely different base assumption: consumers are most likely to shop at retailers who grasp their broader taste profile and can continuously suggest the right products at the right time. You need a fundamental understanding of both the consumer and the product features to find a customer-buyer fit.

Ester is combining expertise in bioscience and artificial intelligence to build the ultimate product recommender system. We are glad to serve the team throughout the whole product development journey as a technical partner with a focus on web development aspects.
We're operating at the intersection of bio and data science. We're glad to know that our web infrastructure and the end-to-end user experience are in the capable hands of the Panenco team.
Jean Van Damme, Founder CEO of Ester
Our joint mission statement
To match customers to the right beer and wine based on their unique flavor preferences and create satisfying consumer experiences that spark flavor exploration.
Product roadmap
Problem statement
How can we optimise the predictive value of traditional recommender systems by leveraging textural and chemical product information from research labs?
Product definition
We needed to define the proper interfaces to visualise the underlying technology and to start validating the early assumptions and data reliability. We built an accessible responsive web interface and REST API.
Minimum viable product
An initial product version was built with a heavy focus on data collection and feedback mechanisms.
Feedback loops
We started to intensely engage with the end-users of the platform to capture their findings and early feedback. The team continues to tweak the underlying algorithm to keep increasing the predictive nature.
With its validated underlying technology, the Ester team is currently actively approaching retailers globally to set up proof of concept trajectories to demonstrate the predictive value of the recommender system in specific product categories such as beers and wines.
Unifying multiple perspectives
Building the right product for the right target audience with a sustainable revenue model
We're operating in a complex setting since the underlying product essentially is an API that determines recommendations for the end- users. Our key challenge is to have close interactions with those users to collect their honest findings and aim to measure the additional value that Ester triggers for them.
Ester is (data) science-driven. The initial focus is heavily centred around proof of value of the underlying algorithms. The business model is continuously challenged based on those outcomes.
We're building a React (responsive) web application with a Node.js REST API on top of Google Cloud Platform infrastructure. The recommendation API is isolated in specific microservices. We foresee to set up additional conversational interfaces.
Based on Flavor Intelligence
ESTER's independent recommendations are always rooted in the most comprehensive chemical and sensory analyses available.
Chemical Analysis
Beer and wine samples are sent to the ESTER lab where over 250 flavor-linked chemical compounds are analyzed.
Sensory Analysis
Samples are assessed by a professional tasting panel and a complete flavor profile is created for each sample.
Packaging and Label Analysis
Label categorization, bottle weight, closure type, branding, and visual factors that influence purchasing decisions are included in each profile.
With a keen eye on the entire user experience
A cross-platform approach for web and mobile interfaces.
For Retailers
Reducing choice anxiety and creating engaging, personalized shopping experiences.
Deliver Flavor Based Recommendations
Offer a trusted, sommelier-level experience that always matches customers to the right beers and wines based on their personal taste, not advertising.
Personalize Marketing Campaigns
Tailor promotions and marketing campaigns to match your customers' individual tastes, turning their errands into delightful shopping experiences.
Optimize Your Beer and Wine Assortments
Use ESTER Flavor Intelligence to optimize your product assortment, adjust pricing, and improve product placement both online and in-store.
For Breweries and Wineries
Finding and engaging with new consumers using real-time, actionable flavor insights and market data.
Get the Most Comprehensive Beer or Wine Analysis Available
Use the most comprehensive chemical, sensory, and branding analyses available to confidently launch and position your products.
Compare Your Beer or Wine to Competitors
Access the ESTER Flavor Platform for a detailed comparison of how your beer stacks up against the competition.
Use Flavor to Connect with Your Fans
Join our growing database of craft breweries and wineries that are recommended to retail purchasing teams and thus connected to new consumers without any additional marketing costs.
Get Technical Alerts
Receive an alert when off-flavors are noticeable in your beer or wine, and get actionable steps to improve flavor quality and consistency.
Making Ester happen
Co-Founder Ester // CEO
Co-Founder Ester // CTO
Co-Founder Ester // COO
Ester // Product Owner
Ester // Marketing Lead
Ester // Data Engineer
Koen Verschooten
Panenco // Product manager
Oleksandr Kanskiy
Panenco // Product designer
Dries Croons
Panenco// Software engineer
Kyryl Kurylchenko
Panenco // Software engineer
Vlad Chernii
Panenco // Software engineer
The only way is up
We're working closely alongside the core Ester team to refine all aspects of the product, from the overall strategy to the user interface and key back-end engineering. We love bold ambitions, so we found a great match with this whole team. Ready to conquer the world!
We're curious to hear about your upcoming product innovations