Marsello is one of the only omnichannel marketing software platforms that combines data-driven automation, email, SMS, and loyalty programs to deliver targeted marketing that boosts customer retention and sales.

Built for retail and e-commerce businesses, Marsello’s approach to data management integrates with point-of-sale and e-commerce platforms and can be used by businesses of all sizes.

Founded in 2014 in Wellington, New Zealand, Marsello empowers more than 5,000 retail and hospitality businesses in over 170 countries.

Businesses using Marsello during the COVID-19 pandemic could quickly pivot to expand e-commerce channels to increase online sales by using features including data-driven automation, emails and SMS campaigns when in-store shopping was reduced or temporarily closed.

With the ongoing rapid growth of its customer base, Marsello’s database is constantly evolving as the business continuously improves its platform. They also need to keep up with increasing data requirements to ensure their sellers maximize sales and customer satisfaction.

As a highly fluid, data-driven marketing platform, Marsello must efficiently expand its feature set and increase its data processing capacity to meet the demands of its customers. To do that the business has to overcome four key technology challenges:

• Incorporate large amounts of dynamic data into its platform
• Introduce new features to the market quickly and iterate at speed
• Enable retailers to build sophisticated segmentation that allows them to segment their customer base in a variety of ways
• Scale to support massive growth, while maintaining high availability for an outstanding customer experience

Marsello was built on Microsoft Azure, so it would have been natural to use SQL Server as a database, but a relational data model would not have allowed for that kind of flexibility.

According to Vinodh Jayawardene, Marsello’s General Manager of Engineering, “the issue with SQL Server is changing the data model because the database needs to be updated if there are schema changes. It just does not perform when you need to make iterations quickly and efficiently,” he said.

As Marsello grew its customer base and platform features, database complexity grew along with it. They wanted a solution that would enable ongoing changes without the challenges associated with constant schema changes.

The Marsello software includes a segmentation function that enables users to select any data attribute such as age or location. With this function, they can put conditions on values and conduct dynamically paired queries. This makes it possible for retailers to segment customers by demographic criteria and buying behavior to maximize sales. The Marsello platform also recommends customer segments to optimize marketing activities, such as email and SMS, based on buying behaviors and other attributes, including recency of their last purchase. This means there are huge numbers of different possible queries a customer could want to make of the database. Offering that at scale and low latency is a particularly tricky engineering challenge.

The challenge for Marsello was how to index this when tens of millions of records exist, which is often the case for Marsello’s largest customers.

They also needed to scale up to handle peak times, such as Black Friday and Cyber Monday. A few weeks before these events, businesses ramp up marketing activities with emails and SMSs to their customers. The Marsello platform captures and analyses data, including open rates, click-through rates and link clicks. This information powers automation activities, such as sending new promotions to customers based on their interaction with promotions and other factors.

In addition to marketing automation and email and SMS campaigns, the database needs to accommodate the platform’s other features, including loyalty programs, marketing lists and feedback surveys.

Marsello chose MongoDB for its ability to manage dynamic data, build complex queries and deliver new features quickly. With a document-oriented database (as opposed to a relational one), a document’s schema is dynamic so developers don’t need to pre-define it and can more easily make changes and incorporate different data types.

This was a game-changer for Marsello in developing its products to offer highly flexible marketing technology solutions. Fields can vary from document to document, and developers can modify the structure at any time, avoiding the disruptive schema migrations that previously frustrated the dev team.

As Vinodh explains, “MongoDB takes it a step further so you can create a model and you can actually push that model with code, and you don’t need to worry about maintaining the schema. That’s a big plus for us.”

To get the most out of the database, their small development team Marsello chose to use MongoDB Atlas, MongoDB's fully managed cloud database service that supports availability, scalability, and compliance with the most demanding data security environments. Marsello had already made the strategic decision to run the business on AWS because, as Vinodh explained it, they wanted to have everything completely catered for them by the experts.

"We are a SaaS provider that uses many SaaS products. MongoDB and AWS's managed services were natural for us to adopt because our thinking is, we shouldn’t have to worry, so let’s get the experts to manage it for us.”

Vinodh Jayawardene, General Manager of Engineering, Marsello

When it came to tackling challenges with segmentation, Marsello realized the value of having a fully managed database on MongoDB Atlas with experts on hand. As Vinodh notes, “We conducted some experiments with the indexing, but that didn’t help us. It wasn't until we worked with MongoDB’s technical support that we really got on the right path."

Another feature of MongoDB that contributes to productivity is MongoDB University, which offers free downloadable video lectures, hands-on labs and quizzes, active discussion forums and proof of completion. MongoDB University offers both Developer and Database Administrator learning paths.

Vinod notes that MongoDB University enables the team to “stay on the same page” as they improve their platform knowledge. This also strengthens Marsello’s employer value proposition, making them more attractive for talented engineers to join.