A Marketplace Development Solution: Making a Review Website

Challenges of Creating a Customer Review Platform

Our client, being a young marketplace company with talented and ambitious employees, wanted to build a review website which could become a single place of truth for customers seeking for any possible products, services or brands. This should be a robust platform with a complex but intuitive structure combining company profiles, customer reviews, top lists, professional articles with advice, as well as the latest news and trends.

The core element in the future system should be an ultimate tool for collecting fresh and honest customer reviews of any possible brands and services. They should be automatically checked, handled, structured, and quickly uploaded to the platform.

In other words, the client needed an in-house review management platform to handle hundreds of reviews per day. Here is the full list of goals the client needed to reach as a result of this review platform project:

• Create the appropriate backend architecture for the future online review platform;
• Build the solution capable to collect, check, structure, and manage customer reviews;
• Enable the tool with robust functionality like automatic speech to text conversion;
• Set the automatic human loops creation for reviews verification;
• Ensure a high level of security and smooth performance of the tool.

For these purposes, the client was looking for highly skilled Python engineers able to set the architecture and build the needed solution from scratch. They also cared much about the possible candidates' English level and time zone, as they wanted them to strengthen their domestic software team and quickly become an integral part of the review website project.

By then, Fively engineers had already successfully worked with the client on creating several custom cloud applications, and received positive feedback. So after a short period of searching for a vendor, they turned to us, showed us all the review website project details and goals, and chose Fively as the best-suiting software company for it. Our collaboration on a common project started in 2021 and lasted for almost a year.

How Fively Created a Review Website Automation Tool

Fively backend engineers worked closely with the client’s in-house team to build a game-changing review website with robust review verification and automation tool. During this project for a website with reviews and ratings, we used the strategy described below.

Project architecture. Our Python specialists built a smart logic for this online review management platform, so that customer reviews are smoothly handled in 2 steps:

Firstly, using Python language and the Django framework, we developed the reviews automation service with an admin panel. The service accepted requests, either written or as phone calls, and created reviews in the database, ready for the verification process.

Then, simultaneously with their creation, the requests were sent to the AWS platform. There, with the help of the a2i service, for each review were created corresponding human loops - these are the scope of tasks for the client’s team to fulfill upon the reviews. The list of tasks included a check for the adequacy of reviews, a check for the right automatic speech-to-text conversion, etc.

After all the tasks upon the reviews were completed, the reviews were structured and put in the right category in the database.

During both these stages, we used Python dynamically typed cross-platform language together with its high-level framework Django. We also used AWS as a database to store customers' data, and its numerous services which helped to build the suitable infrastructure:

• S3 to store and retrieve any amount of reviews 24/7;
• Lambda to perform all the administration of the review automation tool;
• Step Functions for smooth coordination of the components of the tool;
• EventBridge to trigger workflows for further reviews handling;
• a2i for automatic human loops tasks creation.

Team structure. We implemented the full scope of this review platform development project with 3 of our senior backend engineers, who are well-versed in Python frameworks, database management, and are highly experienced in utilizing the AWS toolset.

Communication tools. During this rating platform development, we worked closely with the client’s software team and ensured the top level of transparency upon all the arising questions. We mainly used GitHub for project management and status updates, plus Slack for brief discussions of the tasks. We also provided the client with regular standups and reports to enable clear communication during the review solution development for the client's website.

Marketplace Development Solution Project Results

Due to the chosen smart architecture and the state-of-the-art technical stack, we managed to create a solution that not only automated all the review management for the website but helped to optimize the client’s workflow and boost trust in this platform among customers.

After 12 months of fruitful work, we were ready to present the results of the eCommerce platform review project.

• We created a complex logic and set up an appropriate architecture for the review management platform so that all the requests can be perfectly verified, structured, and published with minimum human engagement.
• We built a robust automation tool for a review website according to the client’s vision and goals. It is capable of handling both written and verbal reviews, analyzing them, structuring, recognizing speech, and turning it into text form.
• Our specialists set the automatic human loops creation for thorough review verification, structuring, and management, which resulted in maximum workflow optimization for the client’s team.
• Smart and quick review handling helped the client automate their workflow, and create a marketplace that became a single point of truth for all-type brands and businesses for clients all over the world. This resulted in a 55% increase in the represented brands' awareness and trust among the platform customers if compared to other user review websites.
• We provided a high level of security for the tool, tuned its performance, and resolved all the technical issues to ensure its flawless and stable work with multiple review types.