With Fivetran, Snowflake, dbt and Looker, Spendesk saves engineering resources, improves the analyst function and introduces a self-service data desk.
With a modern data stack, Spendesk has a scalable solution for growth, improves access to data, frees up 40 hours of engineering time per week, and empowers its centralized analytics team to conduct more proactive and deeper analysis. By creating a self-service data desk, Spendesk has allowed the entire organization to get quick answers to its questions. With reliable access to up-to-date data, an analyst builds a dashboard to determine the business impact of lockdown during COVID-19.
Transformations: dbt (data built tool)
BI Tool: Looker
Spendesk is an all-in-one spend management solution that gives finance teams full control and visibility over company spending. Features include custom approvals, flexible payment methods, automatic reconciliation and integrated accounting. The business is growing quickly, with nearly 200 employees, including a 10-person centralized data team of both data analysts and scientists.
As Spendesk grew and brought on more applications, making decisions based on data became more difficult. Improved access to data would help the business meet certain goals, including:
Measure the performance of every step of the acquisition funnel, from lead generation to upsell
Enrich qualitative product feedback with quantitative data from the platform database and website events to better segment and analyse the customer
Improve business operations with lead scoring, user research and forecasting
Automate product features, including fraud detection, payment reconciliation and image processing
The business tried an alternative pipeline solution and found that the maintenance involved a lot of manual configuration. With every creation of a new table or schema in MySQL, for instance, the team had to reconfigure the connection. It wasn’t scalable.
Spendesk needed a scalable and adaptable solution that could keep up with the business as its quantity of data, data model, teams and processes evolved. It also needed something that was both time- and cost-effective, because, as Damien Maillard, Lead Engineer at Spendesk, explains, “the value of the data team is neither to develop a custom infrastructure nor to spend time maintaining it.”
Fivetran met Maillard’s requirements better than competitor products:
“We chose Fivetran because we believe it is the most robust tool on the market. It performed best at quickly replicating our MySQL, Postgres and MongoDB databases. Fivetran reliably updates data, has a clear view of sync history, and keeps up with evolving data structures, including new tables and schemas. There is no action required on our end, making it scalable for the business.”
Spendesk uses Snowflake as its data lake and dbt to transform the data into a data mart that is queried by Looker for visualization. Snowflake has a number of benefits: low maintenance, cost-effectiveness, and ability to scale effortlessly and precisely in seconds. Spendesk uses dbt to transform the data, which excels in turning "select" statements that reflect core business logic into tables or views, performing all the orchestration and the materialisation. It then uses Looker for its analytics because of the power, speed and controlled iterations of the tool.
Maillard estimates that, without Fivetran, Spendesk would need one full-time engineer to do ELT. Fivetran does more than reduce engineering time, though; it also enhances the job of data analysts:
“We have freed up half of the time of our data analysts that was previously dedicated to ad hoc reports and data extraction. The breadth of data we now have enables the team to work on proactive and deeper analysis. They’re data partners for the other teams in the organization.”
More than ever, Spendesk sees data as a critical response to the current economic environment. “We have to use data to quickly determine what direction we need to go in in today’s market,” says Maillard. “When the lockdown was announced, an analyst was able to quickly put together a dashboard that shows how that could impact Spendesk, capitalizing on previous analysis and metrics in Looker.”
With its current stack, Spendesk has been able to create a self-service data desk that can answer a range of questions from big picture overview, such as “what is the evolution of our Cost Per Acquisition?” or “what are the conversion rates by country or acquisition channel,” to more specific figures, such as “what has been the impact of the feature A on our market fit?” This enables different teams to find their own answers to questions in Looker, while analysts can work on deeper analysis requiring more statistical or data knowledge.
“Every part of our data stack is built to scale and support our fast growth,” says Maillard. “Onboarding people into this data environment is easy, and we have continuous integration and version control for all transformations and visualizations. We are confident that this stack is aligned with the growth of Spendesk.”
About Fivetran: Shaped by the real-world needs of data analysts, Fivetran technology is the smartest, fastest way to replicate your applications, databases, events and files into a high-performance cloud warehouse. Fivetran connectors deploy in minutes, require zero maintenance, and automatically adjust to source changes — so your data team can stop worrying about engineering and focus on driving insights.
About Snowflake: Snowflake is the leading data warehouse built for the cloud. Its unique architecture delivers proven breakthroughs in performance, concurrency and simplicity. For the first time, multiple groups can access petabytes of data at the same time, up to 200 times faster and 10 times less expensive than solutions not built for the cloud. Snowflake is a fully managed service with a pay-as-you-go-model that works on structured and semi-structured data.
About Looker: Looker is a modern platform for data that offers data analytics and business insights to every department at scale, and easily integrates into applications to deliver data directly into the decision-making process.