UntieNots: Boosting the potential of a retail hyper-personalization platform primed for global growth

About UntieNots

UntieNots provides a Software as a Service solution to the retail industry that makes hyper-personalized offers scalable, highly effective, and fun. Using AI, its technology enables retailers to offer a unique proposition to their shoppers and improve the ROI of their promotional campaigns.

Industries: Technology
Location: France

UntieNots built its hyper-personalization and gamification solution on Google Cloud Platform so it could develop innovative models and scale easily to meet demand, fueling its international growth.

Google Cloud Results

  • Facilitates rapid scaling with Google Kubernetes Engine, accommodating millions of additional customers when a new campaign is launched
  • Supports bootstrapping and quick international growth without upfront investment thanks to the global infrastructure of Google Cloud
  • Enables the development of sophisticated personalization models that increase a customer's weekly spend by up to 12%
  • Enables a loyalty program that delivers average monthly savings of €10 to customers

Helps to handle multiple terabytes of retail data for hyper granular personalization

Faced with a fast-paced retail landscape, where customers increasingly value convenience over familiarity, many companies are changing the way they operate loyalty schemes. Rather than one-size-fits-all points or cashback cards, they want to provide a more tailored offering. To help retailers do that, French company UntieNots has developed a Software as a Service solution that uses the power of AI to create hyper-personalized deals that are both playful and highly effective in capturing a bigger share of wallet. Founded in 2015, UntieNots works with retailers, such as Auchan and Carrefour, to target more than 25 million customers yearly worldwide.

"When we founded UntieNots, we chose to bootstrap from the beginning, which required a platform that could grow progressively. Our priorities are being able to handle large volumes of data without significant capital expenditure, and rolling out services internationally."

Zyed Jamoussi, Co-founder and President, UntieNots

"Mass promotions are incredibly inefficient, and result in a poor ROI," explains Zyed Jamoussi, Co-founder and President of UntieNots. "Our Loyalty Challenge software enables retailers to offer one-to-one, targeted promotions that are based on actual customer behavior, such as their preferred brands and spending patterns." Customers access their deals through a retailer-branded web page powered by UntieNots' software. Then they select from several "challenges" that encourage them to focus their spending on their most relevant brands with a retailer, claim discounts, and save money. For example, a customer might have to spend €5 weekly on a kind of pasta over a month to access a discount on that brand. Everything is personalized for each customer: the brand or product, the target amount, and even the discount. By taking on the challenges selected by UntieNots' software, customers save up to €10 monthly. As for retailers, they can expect to see up to 12 percent incremental spend weekly from customers using the platform.

To target customers effectively, UntieNots generates the most relevant offers for each individual by processing large volumes of data that are analyzed using machine learning (ML) algorithms. Looking for a cloud provider that could easily handle significant data volumes and scale to meet the demands of its growing global client base, UntieNots decided on Google Cloud Platform (GCP) as the right platform to meet its needs.

"When we founded UntieNots, we chose to bootstrap from the beginning, which required a platform that could grow progressively," says Zyed. "Our priorities are being able to handle large volumes of data without significant capital expenditure, and rolling out services internationally."

Building an infrastructure that scales effortlessly

Building scalability into a business requires flexibility on a number of fronts: not only handling variable traffic, but also responding to changing client requirements. UntieNots wanted to grow its business without significant upfront investment, so it needed to be able to move fast, both in terms of acquiring new clients and developing its services. "To feed our models and achieve the best predictions and personalizations, we need to process a lot of data on a daily basis," says Cédric Faucheux, Lead Data Scientist at UntieNots. "Because machine learning is a developing area, we also want to test out and implement new ideas in an agile way."

"Google Kubernetes Engine is extremely well integrated with other Google Cloud Platform services, which makes it very intuitive to use. As some clients might only run campaigns for a single season or over a short period, being able to accommodate millions of additional customers at short notice is really important."

Cédric Faucheux, Lead Data Scientist, UntieNots

UntieNots built a microservices infrastructure using Google Kubernetes Engine (GKE) to power its solution, as well as using virtual machines on Compute Engine and Stackdriver for monitoring. Cloud Dataproc makes it easy to manage Hadoop and Spark clusters.

"Google Kubernetes Engine is extremely well integrated with other Google Cloud Platform services, which makes it very intuitive to use," says Cédric. "As some clients might only run campaigns for a single season or over a short period, being able to accommodate millions of additional customers at short notice is really important."

UniteNots uses Cloud Storage both for long-term storage of the data it generates, and for storing clients' data. Each client account is run as a separate project, ensuring data remains separated. Cloud Dataprep helps UntieNots to easily manage multiple terabytes of customer data and reassures them that their data cannot be accessed outside of the organization.

"Increasingly, one of the main ways we access client data is using BigQuery," says Zyed. "As many of our clients use BigQuery, they can simply grant us access without having to move their data anywhere. That's reassuring for them, and less work for us, as it means we can easily comply with local data storage regulations."

Training personalization models with TensorFlow

UntieNots uses Google Cloud AI Platform Training, part of AI Hub, to train its TensorFlow ML models. "As it's serverless and fully hosted, the Google Cloud AI Hub is straightforward to use. We just send the code, and we don't have to worry about what's behind it, so we can focus on developing the best models," says Cédric. With GKE, UntieNots is able to continuously deploy code so updates are automatically integrated.

"When we want to set up a client abroad, in Taiwan for example, we can do it in a matter of days. Without Google Cloud Platform, we would have had to establish a physical presence there first. It's really supported our international growth."

Zyed Jamoussi, Co-founder and President, UntieNots

"Time to market is hugely important for us, so being able to ship code so easily really makes a difference," adds Zyed. "It speeds things up a lot and allows us to roll out new features all the time."

By analyzing transaction data, UntieNots' software is able to identify brands that a customer prefers or is very likely to prefer and offer him a tailored reward for spending more on those brands at a particular retailer.

Supporting further global expansion

With the support of the global GCP network, UntieNots has successfully grown its business internationally over the past four years. "When we want to set up a client abroad, in Taiwan for example, we can do it in a matter of days," says Zyed. "Without Google Cloud Platform, we would have had to establish a physical presence there first. It's really supported our international growth."

UntieNots is also able to keep its latency low by replicating its infrastructure in data centers located close to its clients and making sure its service is available around the clock. "We've had customers take on challenges as early as 1AM, so our solution needs to always be available," says Zyed. "Our uptime is over 99 percent, and we have alerts set up in case anything goes wrong, so we don't need a dedicated team maintaining the servers."

The company is now focused on launching in the United States, one of the world's largest retail markets for promotions. It will also continue improving its algorithms to keep its offers fresh for customers. "We plan to extend our usage of new models in TensorFlow," says Zyed. "That will keep a big part of our team busy over the next few years assessing new models, tuning them to our industry, and rolling them out in the solution."

"Google Cloud Platform has played a big role in our growth internationally," he adds. "Thanks to its simplicity, speed, and reliability, we're now able to focus on improving the service we offer our clients, and expanding to new territories."

About UntieNots

UntieNots provides a Software as a Service solution to the retail industry that makes hyper-personalized offers scalable, highly effective, and fun. Using AI, its technology enables retailers to offer a unique proposition to their shoppers and improve the ROI of their promotional campaigns.

Industries: Technology
Location: France
Google Cloud Platform logo

12 Months FREE TRIAL

Try Kubernetes Engine, BigQuery, and other Cloud Platform products with $300 in free credit and 12 months.

TRY IT FREE
Google Cloud Platform logo

12 Months FREE TRIAL

Try Kubernetes Engine, BigQuery, and other Cloud Platform products with $300 in free credit and 12 months.

TRY IT FREE