From spreadsheets to a data portal using dbt-metrics at Fleetio from Coalesce 2023

Fleetio team members discuss evolving from using a simple spreadsheet to developing their own metrics explorer tool.

“Until we hit [the growth] stage, we were completely fine with that spreadsheet, but when he hit that stage, we started to experience a lot of pains.”

- Jeffery Chen, Director of Data & Analytics at Fleetio

Fleetio team members Jeffrey Chen, Director of Data & Analytics, and Jonathan Hollinger, Senior Data Engineer, discuss the evolution of metrics modernization at Fleetio, a B2B software as a service company focusing on fleet management. They discuss their journey from using a simple spreadsheet for tracking metrics to developing their own metrics explorer tool using Streamlit.

Fleetio's journey from a spreadsheet-based metric system to a more scalable, automated solution

The team at Fleetio shares their journey of modernizing their metric tracking system over the past two years. Starting with a simple spreadsheet system, they ran into scaling issues as the company grew and required more sophisticated metrics tracking.

Despite the initial effectiveness of the spreadsheet system, the Fleetio team found it increasingly inadequate as the company expanded. They needed a more robust solution to keep up with the growing complexity of their business metrics.

"We hit what’s colloquially known as the growth stage of a company, right? And what does dbt say about growth stages?... It's all about creating analytical processes that scale and our spreadsheet did not scale," says Jeffery.

Overcoming the limitations of a spreadsheet with a built-in solution

The team decided to build their own system using Streamlit, a tool that allows them to write data apps in Python. This new system is designed to answer common business questions, provide a simplified way to define metrics, consolidate metric definitions, and standardize visualizations.

They also needed a tool that could handle complex SQL generation, interface with Snowflake (their data warehouse), and run locally for development. In short, they decided to “build something on [their] own and really tailor it to the experience that [their] users are used to.”

Jonathan explains, "The first thing we wanted to do is answer the most common questions… They're doing weekly, and monthly, and quarterly business reviews. They want to understand what's going on about these core business metrics." He adds, "We also wanted to put metrics in context...we want them to see that right alongside the visualizations.”

The new system they built allowed them to define metrics alongside their models without having to build dashboards on top of it. This enabled the team to visualize trends, slice metrics for deeper analysis, and metric definitions over time, all while providing users with business and technical definitions of each metric.

The value of custom-building a metrics tool to meet specific business needs

Despite the challenges and technical demands, the Fleetio team found great value in building their own metrics tool. It gave them complete creative control and allowed them to customize the tool to meet their specific business needs.

Jonathan states, "I think the main reason that we have found this to be valuable is that these metrics are very vital to our business. They're a crucial part of what we're doing, and I think it was not an incredible engineering lift to build this on a tool like Streamlit…It gives us complete creative control to build a data portal that integrates additional features like bringing in lineage information…"

Their custom tool allows the team to explore and understand their metrics in a more in-depth manner. It also provides a centralized place for all their metrics, making it easier for team members to access and understand the data. The tool has become vital to their data management and decision-making processes.

Jonathan and Jeffery’s key insights on scaling metrics

  • Spreadsheets, while simple and effective for a startup phase, do not scale well for growing companies
  • The modern data stack can be utilized for metrics, providing a more scalable and efficient solution
  • Building a custom tool can provide more control and customization to meet specific business needs
  • Streamlit allowed the team to create a data application in pure Python, which fit well with their development workflow
  • The metrics explorer tool Fleetio’s team built provides a standardized visualization of metrics, simplifies defining metrics, and allows for easy exploration and filtering
Related Articles