BLOGAnnounced at MongoDB.local NYC 2024: A recap of all announcements and updates — Learn more >

MongoDB Atlas Empowers Omnichat Deliver Omnichannel Chat Commerce Solutions to Its 5,000 Merchants

image of a man working on a laptop




MongoDB Atlas


Content Management



With teams in Hong Kong SAR, Taiwan, Singapore and Malaysia, Omnichat is a leading omnichannel chat retail commerce solution provider across Asia Pacific. The company was established in 2017 when the founder and CEO, Alan Chan, discovered many merchants lost orders because customers could not make instant enquiries when shopping online. He started with a website live chat solution that evolved into Omnichat.

The company offers several solutions, including:

  • Omnichannel CRM Messaging to manage customers and chats in one place for all channels across websites, Facebook, WhatsApp, Instagram and LINE. The messaging feature can be used to answer questions, recommend products or make sales offers when customers are most engaged.
  • Marketing Automation helps retailers achieve all-round customer acquisition, conversion and retention with marketing automation tools, like abandoned shopping cart reminders, coupon and game modules. The website tracking tools also enable retailers to collect customer behaviour, and design customised promotional messages based on customer preference.
  • Online-Merge-Offline (OMO) Sales pairs with revenue tracking across online and offline channels. No matter if customers complete the purchase at eShop or WhatsApp payment link, Omnichat can calculate the actual transaction amount of customers as well as trace back the revenue contributed by every salesperson. It provides huge convenience for the merchants to calculate the commission and simultaneously motivate the salespersons to deliver better service.

The powerful features and reliability of Omnichat have made it the choice of over 5,000 businesses – from small companies to major international brands, including FILA, OSIM, Logitech, Benefit Cosmetics (LVMH), Venchi, Timberland and Lego.


How to manage and scale with an overgrowing amount of dynamic data

Omnichat was consistently adding new marketing features to its omnichannel messaging solution, generating large amounts of data. Their relational database was unsuited for the large volume of unstructured data they needed to manage, which included 400 million chat records. Initially, all messages and related content were stored in relational database, leading to performance and scalability issues.

The data model was not well defined in the beginning. Louis Yau, Omnichat’s Director of Engineering, noted, “We had limited fields for customer profiles, such as names and email addresses. We realised that we would need to add more information to customer profiles based on the diverse requirements of our clients. For example, with some cosmetics brands, customer attributes include skin types and products customers purchased before.

“We would not be able to achieve this in the previous relational database. Or if we did, we would have some very complex schema designs, which would affect our development speed. It could take over an hour to update when we needed to add more fields and change the database schema, which would block application development.”

To minimize downtime, the team would make schema changes late at night. With more changes required, they realized it would increasingly have a bigger negative effect on application development, requiring more time and hindering customer experience improvements.

With around 34 million customer profile documents, 1.6 billion documents quarterly and 400 million message documents, scalability was another issue with their relational database. They were storing a large amount of data and couldn't efficiently scale horizontally, which led to performance issues.

“Sometimes the index in the relational database could not be utilized when there was a large amount of data, and sometimes it would degrade the latency of our API,” Louis said. We tried to scale up the previous relational database, but the performance could not be improved.”

Search was another challenge that arose as the data levels and requirements grew. Having a separate search engine for text search created operational complexity. Omnichat needed to connect the data from two data sources, which created synchronization issues and additional overhead. The data pipeline maintenance cost was yet another challenge that needed to be addressed.

Louis Yau, Director of Engineering at Omnichat


MongoDB Atlas makes it easy to evolve, avoid downtime and enable native full-text search within the same developer data platform

One of the reasons Omnichat chose MongoDB Atlas is its flexible data schema, which supports a wide variety of data structures. Omnichat can now store complex and unstructured data for a range of businesses with diverse customer profile requirements.

With change-friendly design, downtime has been eliminated when evolving the schema with new data. The document data model of MongoDB Atlas has made it easier for their developers to build and deliver new features.

“We benefit from the document database nature of MongoDB,” Louis said. “There’s no limitation on the data model now, and it’s very flexible.”

MongoDB Atlas supports horizontal scaling via sharding, making it easier to scale.

“When we scale up, there is no downtime,” Louis said. “We can easily change the instance size or add another node in the console.”

MongoDB Atlas also has native search functionality, Atlas Search, built into the platform. Atlas Search integrates the database, search engine, and sync mechanism within Atlas’ unified and fully managed platform. Omnichat was able to easily migrate to Atlas Search in less than two weeks, which has provided many benefits in the form of efficiency gains and ease of use.

“When migrating to Atlas Search, we found that the learning curve of building the search function was quite low,” Louis said. In addition, the data pipeline between two data sources has been eliminated, along with the maintenance cost of the search engine cluster and the data pipeline application server.

Omnichat currently uses Atlas Search on 3 native clusters for customer profile, message, and customer behaviour events data.


Leveraging data with MongoDB Atlas supports complex marketing automation to increase conversion

With the help of MongoDB Atlas, Omnichat is able to leverage data to support its marketing automation and become the leading omnichannel commerce solution provider in Hong Kong SAR, Taiwan, Singapore and Malaysia.

After improving customer experience and solving performance issues, Omnichat has been able to roll out new products, including Omnichannel Chat Commerce Solutions. This all-in-one solution integrates customer service, marketing and sales features to solve retailers’ pain points, including managing several social media channels, and solving ineffective promotion and non-cooperation between online and offline shops.

Omnichat’s solutions help its clients dramatically improve results. For example, the global sportswear brand FILA boosted its digital transformation strategy by adopting Omnichat’s solutions to deliver a complete, multi-faceted shopping experience. By using Omnichat’s Omnichannel Chat Commerce Solutions – which includes sending automated messages via WhatsApp to remind customers of unchecked items in abandoned shopping carts – FILA increased its conversion rate for completed transactions by 3.5 times.

Recently, Omnichat has launched a Social Customer Data Platform (Social CDP), which enables retailers to track the first-party data of customers throughout the customer journey across social media platforms and websites.

After the removal of third-party cookies for Google Chrome and iOS14 in 2024, first-party data will become more important. With more data in the mix, MongoDB Atlas will continue to play a big role in unleashing the power of data.

What will your story be?

MongoDB will help you find the best solution.