Using MongoDB Atlas & Search on Atlas to help physical retailers go digital

INTRODUCTION

Delivering online retail fast

Nextar provides all-in-one POS solutions for small businesses and relies on MongoDB Atlas to provide a robust, flexible database environment. However, when the Covid-19 pandemic struck in early 2020, Nextar customers, most of whom operated strictly in the physical world, needed to retail online to offset lockdown conditions. Nextar needed a fast, accurate search solution and turned to MongoDB Search on Atlas. Retailers now collectively offer hundreds of thousands of products via templated online storefronts, hosted by Nextar.

THE CHALLENGE

Retail during the pandemic

Established in 2003, Brazilian retail pioneer Nextar provides affordable, intuitive, all-in-one POS software, designed to simplify management while giving customers total control over all aspects of their store. It is currently in use in more than 50,000 stores and over 190 countries.

Since 2016, the company has relied on MongoDB to meet its database requirements, migrating to MongoDB Atlas—MongoDB’s modern, multi-cloud database platform—in 2019. The MongoDB Atlas developer data platform is the essential backbone of transactions for each store, enabling them to see inventory and orders while providing timely reports.

For the most part, Nextar customers are smaller retailers, such as local stores, who sell primarily – or exclusively – in the physical world. That meant they had a limited web presence and no online retail capability, so when the Covid-19 pandemic struck in early 2020, there was a major challenge. Mandatory lockdowns forced the closure of these retailers, so Nextar decided to create an e-commerce domain where each of its customers could create their own bespoke shop front.

“Every sale is synced to MongoDB and reflected in real time in the relevant catalog, and it is flexible enough for individual stores to configure whether they want to display out of stock items, for example,” explains Samir Braga, CIO, Nextar.

THE SOLUTION

Simplicity meets rich functionality

Nextar was aware of other search solutions on the market, but they knew there were additional complexities related to syncing data from the database to a bolt-on search engine. They chose MongoDB Search on Atlas because it sits directly on top of their data in Atlas, allowing them to easily add Lucene-based search functionality to an application without having to stand up an additional search engine and sync mechanism. This allowed Nextar to simplify their application architecture and spend more time delivering features that their customers need. In addition to using MongoDB Search on Atlas for their online storefronts, they also decided to use it for catalog search in their suite of apps, and for Nexapp, a system for managing reports about sales and cash flow and new sales APP.

“We needed a search tool that could handle complex searches that aggregated data across multiple dimensions, for example, by time period or product type. MongoDB Search on Atlas can do this without needing additional layers or having to maintain other structure components. It gives us everything we need in one platform,” adds Braga.

“Each shop operates as its own data collection so, for example, if I want to find apples in Shop A, I don’t need to search apples across all the collections, I just need to search in that store,” continues Braga. “Each shop also has its own specific store code so we can perform searches faster.”

Nextar is also implementing MongoDB Search on Atlas with three of their newer applications. Nex on web, the company’s web platform, is a multi-tenant solution that is focused on providing mobility and new possibilities to their customers in a distributed and fast environment — a great fit for MongoDB Search given the data and search performance needs involved. Nex E-Commerce, also known as Catalogo, is a B2C e-commerce solution for Nextar's clients. It allows the products on Nextar’s platform to be available for online orders and purchases quickly and efficiently. The Nextar team uses MongoDB Search on Atlas to search and rank the results according to various criteria so that the end user can quickly find the most relevant products. And finally, Nex on App, the app version of Nextar, is integrated with the entire platform’s ecosystem, and allows the company’s customers to access a wide variety of functionalities in the palm of their hand. Here, MongoDB Search on Atlas is used to assist with search and aggregation of orders and sales.

Additionally, Atlas’s multi-cloud capabilities are crucial for Nextar, as they needed a platform to support their migration from on-premises data storage to the cloud. Nextar utilizes both Amazon Web Services and Google Cloud Platform to host their data in the cloud. “Originally the data was stored on-premises but now we are centralizing this data in the cloud. We believe Atlas is the most suitable platform to support us as we migrate to the cloud completely,” explains Sérgio Trovatti Uetanabaro, CTO at Nextar.

THE RESULTS

Faster and more profitable with no fuss

Now, Nextar customers around the world, whether they offer 60 or 60,000 products, can easily sell online and remain profitable even during the forced closure of their business. For Nextar, the combination of speed and simplicity is a key attraction of MongoDB Search on Atlas.

“After we implemented MongoDB Search on Atlas, we could breathe easy because it can hold more connections without having to maintain other structure components, meaning increased uptime and reduced latency,” concludes Braga. “At the same time, we never need to worry about it; it is one less problem to deal with.”

The Nextar team also utilized MongoDB’s Flex Consulting professional service offering: “We were having difficulty identifying some bottlenecks and overloads of some clusters, and we wanted to ensure we were using industry best practices for Atlas and MongoDB Search,” shares Trovatti Uetanabaro. In order to improve performance and tune costs of NextWeb, MongoDB’s professional services team used a tool called Keyhole to help measure performance of clusters. Specifically, there were a dozen high-use queries using COLLSCAN – four of which were pretty slow (30-120s average time), so the team focused on teaching indexing strategies and explaining how to implement them for these specific queries to improve performance. The professional services team also provided Trovatti Uetanabaro and his team with details about how to understand the various metrics available to analyze their own Search performance, via the Atlas Metrics tab and other tools. As a result, Trovatti Uetanabaro says that “These improvements have led to both reduced service costs and response time,” two tangible positive impacts for Nextar and their customers.

NEXT STEPS

If you want to dig deeper into MongoDB Search on Atlas, spin it up at no cost on the Atlas free tier. You can follow along with reference materials and tutorials in the MongoDB Search documentation using our sample data sets, or load your own data for experimentation within your own development sandbox.