Improving Analytics by Using a Modern Data Stack

A cloud-native data stack equips a construction company with better business intelligence to guide planning and decision-making.
December 17, 2020

Businesses of all sizes desire to harness company data for better decision-making and insights that can change the course of their trajectory. In those companies, data teams that seek to arm company leaders and managers with the ability to make data-driven decisions.

Enter the modern data stack — an affordable and manageable solution to bring together all your company’s data to facilitate reliable analytics for business growth. However, choosing the components and pieces to power your stack can be daunting. In this article, learn from Clayton Hicklin, Director of Technology at Emery Sapp and Sons on his company’s journey to data analytics success with a modern data stack.

Focus on Scaling Infrastructure

As a heavy civil construction firm, Missouri-based Emery Sapp and Sons specializes in building essential pieces of infrastructure that make modern life possible — from roads and bridges to highways and other large public works projects.

Fueled by consistent growth, Emery expects to generate $600 million in revenue this year in a highly competitive industry that operates on thin margins.

But as the fast-growing company’s IT team took a close look at the company’s data infrastructure, the team quickly realized they needed a wrecking ball and a new set of blueprints.

The company’s legacy setup, utilizing an on-premises SQL server and a handful of fixed dashboards, wasn’t flexible enough to scale with a company that has doubled its headcount in the past year and a half, largely through acquisitions.

“Our growth in the last couple of the years, and expected growth in the next couple of years, is really driving a lot of our need for better utilizing data and turning our planning and decision-making into more of a science than an art,” said Clayton Hicklin, director of technology. “We think analytics can help us do that.”

Modern Data Stack: Stages to Decision

Hicklin knew he needed a data analytics setup that would keep his small IT team focused on solving big-picture business problems, not maintaining data infrastructure. At the Modern Data Stack Conference 2020, hosted by Fivetran, he shared the steps he took to get there:

1. Assess the current landscape

The company’s existing traditional data warehouse had been set up by an outside consultant before Hicklin worked there.

“The challenge with having a third party come and implement an environment like that is that, on day one, things might be pristine and working well,” he said. “But if you don’t have the expertise and the ownership and continual investment in that platform, it tends to stagnate,” which is what his team experienced.

2. List out key analytics needs

The team started assessing modern, cloud data tools and quickly wrote a list of must-haves.

Emery’s data team needed to empower other departments with self-service options and high data quality. Most of all, the architecture needed to be agile. A top question, Hicklin said, was, “As we have new data sources come online or if we have new needs as an organization, are we able to fold those into the existing environment?”

Lastly, Emery Sapp and Sons needed an intuitive, mobile-friendly user interface because employees travel frequently and often need to check their phones at an outdoor job site.

3. Understand the universe of data sources

One of Hicklin’s teams requirements was the challenge of data integration. Emery Sapp and Sons has many data sources including:

  • Employee time cards
  • Physical goods such as yards of concrete and gravel used on a project
  • Sales/marketing tracking in Salesforce
  • Internal sources such as HR, safety tracking systems and helpdesk tickets

The team knew they needed an automated data integration component to continually address this data source challenge.

4. Choose the right tools for the job

“As we expanded out the number of data sources and the complexity of the data that we wanted in our warehouse, it quickly became obvious we needed a turnkey tool to take that heavy lifting off of our team’s plates,” he said.

Hicklin’s team ultimately settled on a modern, cloud-based data stack including Google BigQuery, dbt, Fivetran and Looker, which has eliminated the need for custom scripting and day-to-day maintenance.

Selecting a number of specialized tools for each part of the data stack allowed the team to find the best solutions for their business at each step, he said.

With their new data stack in place, his team can handle the company’s growing data needs with minimal maintenance.

5. Build on success

Looking ahead, Hickson is eyeing machine learning and predictive analytics use cases that could, for example, better estimate employee turnover and help Emery better manage its business. In the future, their new data stack could even handle telematics data from construction vehicles, which takes up millions or even billions of rows of data.

The data team has also added a new business analyst and the new tools will help him get up to speed and solve important business problems much faster.

Watch all the Modern Data Stack Conference sessions


Start for free

Join the thousands of companies using Fivetran to centralize and transform their data.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.