Rakuten Saves 60% in Infrastructure Costs with Snowflake

This transcript was automatically generated

So Rakuten rewards is the leading, cashback program in the industry.

We offer cashback for more than forty two hundred stores.

Customer can use our platform to earn cashback for their shopping. From customer experience to customer notification, everything is done using DITA and Snowflake is at the core of it all. We run a lot of, personalized rewards campaign, including targeted cashback for our members and also for our merchants. These are time based campaigns, and we have a budget attached to each and every campaign.

We are we are on a critical mission to make sure that we get maximum out of that budget, both for our customers and also for our merchants. One of the first use cases to go to production was, cashback notification.

As market leaders, we are always in lookout for experiences that can improve our customer satisfaction. Cashback notifications are aimed at notifying the customer when there is higher than normal cashback, for a particular favorite merchant.

And, we could create the signal within Snowflake and use the signal to reach out to the customer using an email or push notification. From a merchant perspective, this is amazing because this will help them to meet their goals, like, whether it be sales or bringing in a new customer. We were able to deploy the cashback notification, product within two weeks. The reason why it happened was we all we need to use was Snowflake's, ML function, which is an anomaly detection function to identify the spikes in cashback.

Traditionally, this kind of a deployment or a product would have taken almost eight weeks. So we saved almost two x time to hit the market and also, like, the infrastructure cost. We saved around sixty percent because there was no additional requirement for from an info perspective. This was all done within Snowflake, scheduled within Snowflake.

So, yeah, that was amazing. So usually, an application like this, especially an ML product, requires a specialized data scientist or an ML engineer. But in this case, this was all done by a data analyst with just a SQL background. I'll say we definitely reduce the pressure on our data scientist and ML engineers, thereby giving an opportunity for analysts to be a part of this space.

Snowflake is critical for our enterprise AI strategy. We have been on the lookout for tools that can help us democratize AI within the company. Snowflake Cortex is definitely helping with that.