Overview

pLeader in the women’s footwear, handbags and accessories market in Brazil, ArezzoCo’s data lake runs on AWS using Amazon Redshift solution to analyze consumer data collected over the past three years, which is stored on Amazon S3. With the support of APN business partner Eleflow, ArezzoCo reached 91% assertiveness on items in stock, vs. 80% of human process. In addition, Arezzo reached savings of 7,000 hours of work. a href="https://aws.amazon.com/pt/solutions/case-studies/arezzo/"Read the full case (in Portuguese)./a/p

Supernova Award Category

AI and Augmented Humanity

The Problem

pFive years ago, Arezzo Co started through e-commerce its digital transformation process, with the goal of making this channel more data-oriented. In this process, Arezzo Co identified that data was the main obstacle to decision making not only in e-commerce but also in other areas of the company: in addition to the time needed to obtained, the data was often inconsistent or incorrect. Another problem was the price lists. As the systems were independent and did not talk to each other, the tables had to be loaded in all of them. To unify all this information, Arezzo Co opted to create a cloud data lake, which would also allow it to be scalable throughout the company's maturation and expansion process./p

The Solution

pWith the support of APN business partner Eleflow, Arezzo Co started the process of building its data lake in October 2018. It was born integrated with the Tableau Server solution and initially received the information of greatest interest to the company: sales by categories, brands, channels, customer, etc. After this, the Arezzo Co BI team began to build data views for the areas that were most interested in the information, making it possible to build and access real-time reports. Initially, the platform was used by 120 users, but this number has been expanded, and should reach more than a thousand already in 2020./p

The results

pToday, Arezzo Co's data lake runs on AWS, using the Amazon Redshift solution to analyze consumer data accumulated over the past three years, stored in Amazon Simple Storage Service. The company also uses Amazon Machine Learning algorithms to classify the shoes it manufactures (SKUs) and Amazon SageMaker solution to register these products./p

Metrics

pLabeling shoes using Amazon Sagemaker provides an efficiency of 94%, compared to the manual rating of 80-85%. By running its analytical they´re able to classify the best assertiveness models with the market and identify potential best sellers./p

The Technology

pAnalytics, Data Lakes, Machine Learning Artificial Intelligence - Amazon Machine Learning, Amazon Redshift, Amazon SageMaker Ground Truth, AWS Lambda/p

Disruptive Factor

pAgility Performance, Innovation, Staff Productivity/p

Shining Moment

pArezzo runs its big-data warehouse on AWS, using Amazon Redshift to analyze consumer data of the last 2 years. The company uses machine learning algorithms to label all the 36 thousand SKUs they prototype per year./p