relevanC: Stacking up ROI with a platform that helps retailers and CPGs personalize offers for customers

About relevanC

Founded in 2017, relevanC is a rich data solution provider comprised of two branches: relevanC Advertising and relevanC Retail Tech (formerly Maxit). Focused on the retail sector, its goal is to improve media and marketing ROI through proprietary shopper data, AI, and digital technology solutions.

Industries: Retail & Consumer Goods
Location: France

relevanC Retail Tech built its smart coupon personalization platform on Google Cloud, integrating and querying millions of transactions every day to increase ROI by an average of five times for its clients.

Google Cloud results

  • Offers a self-service platform to make personalized coupon creation easy for retailers and CPGs
  • Scales computational power automatically with Google Kubernetes Engine, optimizing costs
  • Ingests and processes 10+ million daily transactions per retailer
  • Delivers 30+ million weekly personalized coupons for their first three clients

Sales multiplied by three, thanks to offer personalization

According to a report by McKinsey, personalization will be the prime driver of marketing success within the next five years. Because personalization uses data to better understand what customers want, it can be an effective way of delivering content that’s more relevant to their needs, while enabling retailers to spend their marketing budgets where it matters most. By harnessing information on millions of anonymized transactions from some of France’s largest retailers, smart marketing company relevanC is helping retailers understand their customers better and deliver more relevant marketing content.

To do that, its Retail Tech branch provides a suite of loyalty and targeting tools, including a smart loyalty application and a couponing platform. “Our vision is to make personalization easier and more accessible,” explains Benjamin Karoubi, Head of Data at relevanC Retail Tech. “We aim to make it simpler for retailers to gain insights that help them serve customers better. By combining our products, they can carefully monitor ROI and uplift, and optimize their impact.”

“We were convinced BigQuery would be an asset because the big challenge for us is managing the volumes of transactional data that we deal with. We currently work with three major retailers, and for each retailer, we collect transactional data on 10 million products sold every day.”

Benjamin Karoubi, Head of Data, relevanC Retail Tech

After launching the Casino Max, Jumbo Score Max, and Leader Price loyalty applications, relevanC Retail Tech decided to develop a smart coupon personalization platform that could be used by both suppliers and retailers to create offers. “By using our algorithm to create offers that reward loyal customers or incentivize less active customers to return to a particular brand, retailers can multiply the ROI of their campaigns fivefold,” explains Benjamin.

To build a scalable solution that can process millions of transactions every day, relevanC Retail Tech turned to Google Cloud.

“We were convinced BigQuery would be an asset because the big challenge for us is managing the volumes of transactional data that we deal with,” says Benjamin. “We currently work with three major retailers, and for each retailer, we collect transactional data on 10 million products sold every day.”

“When the task is finished, Google Kubernetes Engine shuts down the VMs on Compute Engine automatically, so we only pay for the computation time we use. That’s important for us, as we only need to carry a dozen or so computational tasks a week for generating promotions.”

Benjamin Karoubi, Head of Data, relevanC Retail Tech

Building a coupon personalization platform that’s easy to scale

Scalability should deliver flexibility both in the present and for the future. For its smart coupon personalization platform, relevanC wanted an infrastructure capable of ingesting large volumes of data daily, as well as powering complex computations several times a week. It also wanted to prepare for future growth. “We anticipate more and more brands and retailers joining the platform in 2020,” says Benjamin. “We need our data ingestion to happen automatically, so we can focus on problem solving and improving our personalization algorithm.”

relevanC Retail Tech started working on the platform in September 2018, and by January 2019, the first coupons were already being sent out to retail customers. “The development process went smoothly, because there’s a wealth of rich documentation available on Google Cloud,” says Benjamin. “It was easy for our developers to find answers to their questions and implement the solutions.”

The infrastructure consists of a front end, a back end, and a computational engine. Pub/Sub connects the services, while campaign and targeting files are saved in Cloud Storage. To load them to their IT systems, clients are able to download those files directly from the platform or it can be sent through an API. Cloud SQL stores the range of characteristics that define offers, such as offer duration, budget, and KPIs.

To scale up virtual machines on Compute Engine when there are computational tasks to be completed, relevanC leverages Google Kubernetes Engine. “When the task is finished, Google Kubernetes Engine shuts down the VMs on Compute Engine automatically, so we only pay for the computation time we use,” says Benjamin. “That’s important for us, as we only need to carry out a dozen or so computational tasks a week for generating promotions.”

Ingesting and storing very large volumes of transactions every day

To store and query transactional data, relevanC leverages BigQuery, using Cloud Composer to connect to the retailers’ information systems. “Google Cloud Composer is great because it makes it easy to monitor the jobs that have to be carried out every day,” says Benjamin. “We don’t have to think about it, and if there’s an issue, our data engineers receive a notification and can manage it.”

The KPIs that are recalculated based on the previous day’s data are then made available to the retailer or supplier through the frontend portal of the platform, so they can see exactly how the offer is performing. They can track how many coupons have been sent out and how many customers have used the offer, as well as the uplift created, thanks to an A/B testing function that compares performance without the offer.

“With our infrastructure, we’re able to get new retailers onboarded in a couple of weeks. That’s because of the simplicity of Google Cloud and its ability to interface with any information system.”

Benjamin Karoubi, Head of Data, relevanC Retail Tech

Delivering more than 30 million personalized coupons every week

With the relevanC coupon personalization platform deployed on Google Cloud, more than 30 million personalized coupons are sent to more than 2 million shoppers on the Casino Max and Franprix apps every week. “Customers are three times more likely to use our custom-generated offers than generic offers,” says Benjamin.

Benjamin believes the decision to build the platform on Google Cloud has contributed to its success. “With our infrastructure, we’re able to get new retailers onboarded in a couple of weeks,” says Benjamin. “That’s because of the simplicity of Google Cloud and its ability to interface with any information system.”

As relevanC provides other solutions, the plan is now to expand the use of Google Cloud to other areas of the company. “One of our projects for 2020 is to standardize our cloud systems,” says Benjamin. “We want to integrate our infrastructures on Google Cloud and create more synergy between different teams.” As part of that, he sees relevanC using more Google Cloud services such as Dataproc to have managed Spark services instead of using and maintaining a Hadoop cluster on which some data science teams are currently working.

“Thanks to Google Cloud, we have an infrastructure that can autoscale to meet our needs and optimize our costs,” says Benjamin. “We look forward to welcoming more retailers to the platform in 2020.”

About relevanC

Founded in 2017, relevanC is a rich data solution provider comprised of two branches: relevanC Advertising and relevanC Retail Tech (formerly Maxit). Focused on the retail sector, its goal is to improve media and marketing ROI through proprietary shopper data, AI, and digital technology solutions.

Industries: Retail & Consumer Goods
Location: France