AvaTrade sees over $100k a year in ROI with a data infrastructure cloud migration

Key Takeaways

• AvaTrade migrated its on-prem SQL Server into Snowflake, saving the company $100,000 on database administration and DevOps maintenance resources.
• The new cloud infrastructure using Rivery, Snowflake, and Tableau Cloud unlocked the ability to deliver additional insights to optimize customer engagement and respond to new analytics requests up to 80% faster. 
• With the help of an experienced partner – Vision.bi, the migration of the on-premises data warehouse built over 8 years was completed successfully within a few months and ahead of schedule.
• Rivery provides AvaTrade’s data team end-to-end control over their data integration processes from running automated data pipelines, to orchestrating Tableau processes as well as Snowflake costs management.  
• ROI for the system in the next 12-18 months is estimated to potentially reach hundreds of thousands of dollars in ROI in terms of cost savings, increased productivity, and improved sales.

AvaTrade has pioneered online trading since 2006 being one of the most secure brokers in the industry. It empowers over 400,000 registered customers globally to invest and trade, executing more than two million trades a month. As a B2C FinTech organization, AvaTrade relies heavily on data to provide a superior consumer experience. Since 2013, the business has maintained most of its analytical data on-premises using a SQL Server database. To quickly generate cost-effective and additional comprehensive business insights, AvaTrade decided it was time to migrate all of its data to the cloud.

Building Upon an Extensive Data Platform Leads to Cloud Migration Challenges

AvaTrade had built a lot of infrastructure for their on-premises data architecture. This included SQL Server as the data warehouse, SQL Server Integration Services (SSIS) as the primary ETL tool, and Tableau Server for visual analytics. While the SQL Server data model was well defined and the Tableau dashboards were adopted across the business, the data team spent time when it came to delivering on new business requests. The main challenge was the limited computing power of the on-premises infrastructure. Even though the server had massive hardware behind it with lots of RAM and DBAs spent lots of effort to optimize the environment, data integration was challenging. For Shadi Jaber, Head of Intelligence & Data Science at AvaTrade, this challenge was the first trigger to look at a cloud migration.

While handling projects that required handling huge amounts of data, we kept on spending time on optimization projects such as data archiving which occupied a lot of time.

Buying more cores just increased the costs but didn’t solve the problem. An external consulting team started planning the migration but Shadi noticed there were too many complexities involved in the suggested path forward. It was decided to keep simplicity as a guiding principle for the migration. After a few weeks of planning, AvaTrade determined that the current data warehouse model is worth keeping, meaning there would be little impact on the Tableau dashboard layer. Attention was turned to bringing on some experts’ help to accelerate the data warehouse migration and ensure best practices are being followed.

Having worked with Vision.bi in the past and knowing their reputation in the BI space, I brought them on to help us figure out how to build out our modern cloud data platform.

Planning the Migration

Vision. bi being a top partner of Rivery, Snowflake, and Tableau with deep expertise in data migrations was able to fine-tune the plan to minimize the concerns involved in such an important project. Using their migration assessment methodology, creating a distinction between processes to migrate and processes to redesign, they formed a project plan. At a high level, they took the approach of creating a minimum viable product (MVP) for a few key end-to-end processes including many data sources and respective data models. After the MVP delivery, the plan was to add more models and complete the Tableau cloud migration as well. At the micro-level, they categorized tasks using a scale of complexities from 1-5 and dependencies. They first tackled the migration of the data ingestion processes from source to target and then adjusted the data transformation logic on top of it. This covered 95% of the migration validation with the final 5% done directly in Tableau where it’s easier to get the business acceptance. 

Moving Data From On-Prem to the Cloud

AvaTrade chose Snowflake as its warehouse and stayed on Tableau but moved it from Tableau Server to Tableau Cloud. Since they already used Rivery and liked the experience, Rivery was a natural candidate for handling the data migration to Snowflake. The main question at that point was if they should bring on a separate transformation tool or continue with Rivery for transformations as well. 

We started looking at which ETL tools were innovative enough to handle this migration with the simplicity we looked for. Realizing a migration project is already a complex undertaking, we wanted to avoid having the team learn and maintain yet another tool. Rivery’s integrated push-down SQL transformation and orchestration layer was a great fit for the skillset of my team that was used to writing SQL queries.

In addition to having the ingestion and transformation managed from one place, AvaTrade was also able to use Rivery to gain complete control over all processes from source to dashboard including the orchestration of the Tableau Dashboards datasets extract refresh. With Rivery triggering Tableau extract refreshes right after the data was made ready in Snowflake, AvaTrade was able to ensure business users are always seeing the latest data available in the warehouse without incurring additional Snowflake compute costs for unnecessary Tableau extract refresh jobs.   

With a streamlined architecture of Rivery, Snowflake, and Tableau Cloud, AvaTrade found the simplicity, flexibility, and scalability they wanted to gain with their cloud migration.

With Vision. bi’s project resources, they formed a relatively small and focused migration team that started working on migrating SSIS jobs to Rivery, managing to accomplish the bulk of the migration ahead of schedule. Rivery’s native ability to manage incremental loading processes and automated DDL query generation for tasks such as table creation, saved a lot of time and eliminated lots of SQL code previously managed in SQL Server stored procedures. All in all, this showcased that with Rivery and Snowflake, AvaTrade can smoothly replace mature systems that were built for years without having to use workarounds or spend years migrating.     

Besides connecting to SQL Server and Marketing Ad sources, AvaTrade utilizes connectors to other SaaS applications including Zendesk, SurveyMonkey, and Outbrain – providing richer insights as data analysts can combine data sets to make better decisions.

Moving to the cloud to eliminate infrastructure management was a critical decision for the business’ future scalability. “It also brings the scalability that traditional on-prem setups cannot attain without significant cost.”

Delivering on New Business Requirements at Scale 

Rivery and Snowflake do a lot of the optimization involved in SQL Server for you. The ability to scale up and down dynamically and the ability to break out load compute from transformation compute, from consumption compute – there is infinite value in that.

With the cloud consumption models, AvaTrade pays for what they use and was able to replace the costs of SQL Server with the combined cost of Rivery and Snowflake while eliminating the need for a full-time DBA on guard. 

With a scalable data infrastructure in place, Standard ongoing data integration requests that used to take a week are now completed within one day, and analyzing customer engagement with over 1 million calls and email messages per day, can now be easily done with any dimension the marketing team wants to filter by to find specific message level insights. 

Finally, as the data team at AvaTrade took full ownership, they built a few dedicated data management monitoring dashboards so they could stay ahead of the business and be proactive if any issue arose. 

Migrating your data from on-premise to the cloud? Building your first Cloud Data Warehouse? Rivery can help make this process seamless. Request a personalized demo to learn more or get started today with a free trial.