The Use of AWS at Josera

Problem Statement

Hay is the staple food for horses, supplemented with concentrated feed and additives, such as those available at Josera.de, to ensure optimal care. However, horse owners often struggle to determine the specific dietary needs of their horses and the appropriate products to meet these needs. Each horse’s nutritional requirements vary depending on breed, age, gender, and activity level, while hay, as a natural product, has no fixed nutritional values. The exact nutrient composition of hay can only be determined by lab testing, a laborious process for horse owners. Furthermore, the results often consist of complex scientific data, making it challenging to derive the ideal feed composition for a specific horse. To address this, Josera.de offers a hay analysis service with a straightforward ordering process. Horse owners not only receive their hay’s nutritional values but also an interpretation of the results tailored to their horse, along with specific product recommendations for optimal care.

The main challenge in implementing the backend processes was creating a solution that seamlessly connected the online and offline worlds. Customers needed a simple way to collect hay samples and link them to their horse’s information before sending them to the lab. The lab needed an efficient system to transmit results, which had to be processed and combined with product recommendations, then delivered to customers in an automated fashion. As this hay analysis service was a novel market offering, demand was difficult to predict. Therefore, the AWS solution had to be elastic, cost-effective, and integrate smoothly with the existing AWS infrastructure.

Solution Approach and Architecture

To provide customers with a seamless experience, the online shop was enhanced with a questionnaire and an optimized ordering process, including materials for shipping hay samples to the lab. Since the labs did not have APIs or automated data exchange systems for result transmission, an email inbox using Amazon SES (Simple Email Service) was employed.

The remainder of the solution was then implemented on AWS. The predominantly serverless architecture is illustrated in the following diagram:

When a customer orders a hay analysis, their questionnaire responses and order details are extracted via the shop API using AWS Glue and stored in an Amazon Redshift-based data warehouse for further processing and analysis. When the lab sends results via email, they are received through Amazon SES, which triggers an AWS Lambda function to save the message in an Amazon S3 bucket. Once the message is stored in S3, another Lambda function extracts the lab results and archives them in the data warehouse. The order and horse information are then combined with the lab results, processed using Python libraries to generate a detailed analysis and tailored product recommendations in PDF format, which are stored in an S3 bucket. Josera’s experts can update their knowledge on result interpretation and product recommendations through a decision matrix stored in an S3 bucket. Finally, the results document is sent to the customer via SES after being retrieved from the S3 bucket.

Outcome

This solution significantly simplifies the hay analysis process, providing horse owners with an easy way to optimize their horse’s nutrition.

For foodforplanet, the process is almost entirely automated in the AWS cloud, eliminating manual effort after the hay analysis is ordered. Analysis results are fully automated, saving 30 minutes of manual effort per lab result and ensuring customers receive their results within 1-2 minutes of the lab’s email.

b.telligent – A Reliable Partner in the AWS Cloud

b.telligent is an independent consultancy specializing in optimizing digital and data-driven business processes, customer, and supplier relationships. As an AWS Advanced Consulting Partner, b.telligent has over 300 employees across ten locations in Germany, Austria, Romania, and Switzerland. The company was tasked with designing and implementing the backend solution for the hay analysis on AWS. b.telligent has supported Erbacher the food family for several years in building their AWS cloud data platform.