Leading tech company sees 5x faster search speeds with MongoDB Search on Atlas

INTRODUCTION

Helping the world to work better

A leading technology company is dedicated to developing solutions that enhance and streamline various industries’ efficiency.

“We have a huge focus on innovation. It’s not just our job, it’s our passion. We help companies to operate faster for less money, be more sustainable and improve quality,” says the company’s Principal Solution Architect.

Today, the company is focused on optimizing the construction industry division, which provides digital solutions such as project management tools, apps to help architects and engineers with modeling and designs, and workflow automation for contractors.

THE CHALLENGE

Helping estimators find the right components, faster

The construction division’s content platform helps users to create estimates and determine the volume and costs of a specific product or material required to support a building design, including components such as light switches, wiring, and pipes for new buildings. This is part of a construction workflow, which connects and streamlines design, estimating, bids, and procurement.

The properties of these 800,000 components can vary greatly. The content platform needs accurate information such as the size, material, and shape of every component, but data is compiled from hundreds of third-party manufacturers who all track different specifications and display their data differently. The company uses JSON format to store and transmit this diverse data in jagged arrays.

To improve discoverability of components and the relevancy of search results, the team needed to migrate from its relational database to a more flexible solution. “When creating an estimate, contractors don’t want endless options. We needed a simple text search that returned a small list of relevant results based on accurate component properties,” explains the Principal Solution Architect.

THE SOLUTION

Tailoring more accurate and relevant search results

The technology and services company has used MongoDB across the organization since it migrated from SQL Server in 2011. It upgraded from MongoDB Community to MongoDB Atlas in 2019 and rolled out MongoDB Search in 2020. The organization has a multi-cloud environment, but the construction content platform services team hosts the solution on Microsoft Azure.

“MongoDB is the heartbeat of our business. It’s a great solution that aligns with our philosophy of innovation. We did a side-by-side test with another solution on the market and the performance and efficiency was mind blowing,” says the Principal Solution Architect. “MongoDB Atlas has so many game-changing features, we were excited to see how we could use it to get better search results.”

The team used MongoDB University to learn how to work with a NoSQL database and engaged the MongoDB Flex Consulting team for advice on how to define an index with a complex JSON structure. They built an aggregation pipeline query for simple text searches and used wildcards to search across multiple component fields. Based on user requirements, they also enabled fuzzy search options, but this was still generating too many results and wasn’t creating the consumer-like experience the team was hoping for.

The solution to improving the impact and efficiency of search results came down to facets. “I had a eureka moment when the MongoDB team explained that aggregation facets are not the same as Search Facets,” reveals the Principal Solution Architect. “With the proper use of Search Facets performance went through the roof and we really refined our search results.”

This included adding drill-down classifications and optional filters. Using keywords was initially bringing up too many diverse results, especially when entering catalog numbers including hyphens. The team created a custom analyzer to keep text in a single token and updated the search index to resolve this issue.

THE RESULTS

Five times faster searches and more relevant results

The company’s new construction content platform is already seeing impressive results. Searches are five times faster than previously, and it’s not just the speed, but also the quality that has improved. “We went from getting 10,000 search results to 100 more relevant hits, and search aggregation has improved by 80%,” says the Principal Solution Architect. And more productive staff are more cost-effective, as they can respond to five times as many requests for estimates than previously and the platform will scale to support up to 30-50 million components. The developer team is also more efficient with MongoDB. The company has implemented standard best practices that help them get new releases to market faster. “We’ve amped up our toolbox ten-fold with MongoDB Atlas. For example, our DevOps teams aren’t spending hours doing cluster migrations, I can do it myself in 15 minutes,” says the Principal Solution Architect. “This is next-generation efficiency.”

The wide-scale adoption across various teams means there are endless opportunities for innovation — graphs developed on MongoDB for the agriculture division, for example, could be used to develop solutions for the transport or building monitoring departments.

“MongoDB Atlas is not just a database. I love it, and I love that it keeps on innovating,” explains the Principal Solution Architect. “We’re looking forward to additional Atlas features such as search autocomplete, better dollar lookup, Atlas GraphQL API, and Atlas Charts to support content management and further improve performance.”