Cash Converters controls costs, accelerates access to insights with Azure Synapse Analytics

Cash Converters is Australia’s largest second-hand retailer and also offers its customers a broad range of personal financial services. The company has expanded internationally and now has more than 700 stores in 16 countries and a thriving online business.

Established in 1984, the organisation also has a philosophy of continuous improvement across all areas – including IT and data science.

That focus on continuous improvement led Dr Ryan Behdad, the company’s data science lead, to champion the deployment of Azure Synapse Analytics. Previously Cash Converters used Azure SQL to build and manage its data warehouse, but the opportunity to migrate to Azure Synapse Analytics and take advantage of the pay-as-you-go features was compelling – as were the additional features and functionality.

The transformation was completed in just a few weeks and according to Behdad, Azure Synapse Analytics proved simple to deploy, configure and use.

Cash Converters has data galore – from its retail outlets, online sites and operational systems. Customers can buy online, get quotes for items they’d like to sell, and apply for or manage loans through the website or Cash Converters app – with data collected at every step. This is designed to deliver the enterprise-wide transparency that managers and senior executives need to run the business efficiently.

Behdad explains; “We have a lot of data available and we always want to make the best use of it. It enables us to make data informed business decisions.”

The data details Cash Converters’ 200,000 personal finance customers as well as the 1.5 million store transactions that take place each year – the entire data collection can reach up to 7 terabytes.

The insight from data allows managers to perform credit risk assessments on applications for personal finance, and predict the likelihood of the customer making regular loan repayments if the application is granted. The data ingested by Synapse comes from a data extract refreshed on a nightly basis to ensure managers can make decisions on the most current information available.

According to Behdad this approach benefits both the business and customers already and adds that; “Using machine learning for retail is on our roadmap.” That would, for example help Cash Converters understand where to best locate its stock, based on customer demand, and also where there are opportunities to open additional outlets.

The organisation uses data operationally – to manage the buying and selling of goods as well as to handle loan origination and ongoing loan management.

Its analytics focus is on interpreting that data to make better decisions and understand how the business and the market is performing, and what customers are seeking.

It also uses machine learning to run predictive modelling that can be used to shape strategy for the longer term. This, for example, reveals behaviour patterns across different cohorts that can inform Cash Converters about the next products it could develop to address customer needs.

All three use cases were the focus of the Azure Synapse Analytics transformation says Behdad.

Speedy transformation

The decision to transition from Azure SQL Data Warehouse and Azure Databricks to Azure Synapse Analytics was fairly straightforward according to Waqas Hussain, an experienced data engineer in Behdad’s team. While the team was aware of other technologies in the market, such as Snowflake, the opportunity to remain within the Azure ecosystem promised a smoother, faster and more cost-effective transformation journey.

Change management was also minimal as the IT team is familiar with the Azure stack, and business users found the transition to the new platform straightforward. “They can keep using familiar stuff, for example Microsoft SQL Management Studio to access the data. Everything is readily available,” adds Hussain.

Azure Synapse Analytics also promised more value. “You don’t have to pay any upfront cost. You just pay for what you are actually using,” he says.

Cash Converters is already saving more than 40 per cent because of the transformation, says Behdad, because there is no longer the need to run a dedicated cluster 24×7 – the company just pays for what it uses.

Because Azure Synapse Analytics is integrated with Azure Active Directory which is already used across Cash Converters, access to the Synapse system is easily managed and controlled.

In the first instance, three of Cash Converters’ data sources have been transitioned to Azure Synapse Analytics making the data very accessible according to Behdad. This data is in the personal finance domain, and is mainly used for querying, reporting and dashboarding. “We can simply write SQL queries in any programming language or in any environment. We can have a Python script that just uses a SQL query and connect through ODBC to the data sources and have access to the data without needing to install special libraries.”

He adds that integrating that with the Redash open source visualisation tool, that the company uses has also been straightforward.  “It was really simple, just adding connections to Synapse and the data was immediately available in the visualisation and lots of different charts and dashboards were created based on that data.

“All the pipelining behind the scene was really fast. It was really quick and it was very efficient.”

Now, people across the business, ranging from analysts to senior managers have easier access to information in the forms of dashboards and reports built on top of this data.

Besides the core group of data scientists supporting the organisation as a whole Behdad’s approach has been to have data scientists spread out across Cash Converters to support different business units with their analytics needs day by day. He’s now focussed on enhancing the quality of the data available to the business units, and is exploring whether the Azure Synapse Analytics Workspace could accelerate access to insight – all in the spirit of continuous improvement.