Scaling data in construction with Emery Sapp & Sons
Oct 26, 2020
Emery Sapp & Sons is an employee-owned heavy civil engineering company based in Columbia, Missouri. We focus on complex private and public sector projects ranging from excavation, grading, underground utilities, and bridge construction to asphalt and concrete paving. Since our founding in 1972, we’ve grown to over 1,400 employee-owners, due to organic growth as well as acquisitions.
Our data has scaled right alongside our company. We analyze information from our administrative, financial, and human resources departments, along with field data from job sites and telemetry equipment. We know that leveraging this data intelligently is the best way to successfully scale, remain competitive and nimble, and provide the best service to our clients and employee-owners.
Choosing a modern data platform with Google Cloud
Previously, we primarily relied on Excel workbooks for data analysis. However, with the growth of our business (and of our data volume and complexity), we knew this wouldn’t continue to scale. Excel required too much manual work, and increased the risk for human error. A foray into other workbook-based BI tools resulted in inconsistent data — when our accounting team needed to know how many hours someone worked, they got different answers depending on the tool they used and where that tool pulled data from. That wasn’t going to cut it for us.
It was time to reinvest in our data. Not only did we do this by finding the tools best suited to our needs, but also through focusing on our people and processes. After all, a great data tool doesn’t do you much good if no one uses it.
We ultimately chose Google Cloud’s BigQuery and Looker as key components of our data stack. Here’s why:
First, the cloud-based architecture was a natural fit for us. We have a wide variety of data sources, including SQL server-based applications, older databases, Salesforce, and cloud-based data tools. Unifying these sources would give us a holistic view of the business. Additionally, our corporate goals revolve around construction, not IT, so it was important that we didn’t have to build a team to babysit our data infrastructure.
Second, Looker met our need to enable user-friendly self-service across the organization. After the data models are defined in LookML (Looker’s Git-versioned modeling layer), it’s easy for all our employees to analyze, explore, and visualize data. With LookML, I trust that the modeling and data governance happening behind the scenes is consistent, so everyone can confidently access and act on accurate data. And it’s all done through a business-user friendly interface.
Finally, the combination of BigQuery and Looker enables us to control our own data stack with minimal maintenance. Our source data and systems are a combination of on-premise SQL server and PostgreSQL databases and SaaS based services such as Salesforce. We use Fivetran automated data pipelines to centralize both our on-premise and SaaS application data which was easy to set up. All the elements of the stack are integrated but can run independently. This flexibility and control was important to us. We didn’t want to be faced with vendor lock-in down the road and we didn’t want to invest additional internal resources on implementation and maintenance. Previously, I had to spend one day per week building and maintaining custom pipeline automation scripts and doing manual data manipulation; today it’s all automated. This architecture saved us a significant amount of implementation time — and we don’t need a dedicated engineer to maintain it.
Showing unified dashboards despite disparate data
The ability to unify information coming in from multiple data sources onto our dashboards is a crucial advantage of our data stack. Because we’ve made numerous acquisitions over the last few years, we’re in the process of consolidating multiple accounting systems. Until that consolidation is complete, we have to show information (such as billings, costs, time, and employee information) for our customer jobs from those systems to our stakeholders in the most succinct way possible.
Each of the accounting systems defines jobs in a different way. Standardizing dashboards on KPIs that rely on these business terms would be a nightmare without Fivetran, BigQuery and Looker. We use Fivetran to load all the relevant data into BigQuery, and we map the information from the different accounting systems into a common definition of a job. The information might be coming from a dozen tables in those systems with different definitions; but with Looker, we can show a unified view of the data. We can then look at the jobs against our budgets or timelines and understand the business impact.
Eliminating manual reporting and saving time
We’ve realized significant time savings by automating manual reporting wherever we can. One of our accounts payable specialists used to spend six hours every month manually generating cost reports for each of our branches. We’ve now prebuilt the reports in Looker and set up the necessary data feeds, so this data is automatically delivered via email each month. Those hours are now dedicated to revenue driving activities instead.
Another report that has been significantly improved — both in terms of quality and time savings — is a weekly profitability and accounts receivable dashboard. We used to take over two hours to run the dashboard every Saturday so we could present it to our CFO on Monday mornings. Even putting in work over the weekend, that data was already outdated by the time we presented on Monday. Now, Looker automatically generates it so our CFO can look at it every morning with real time data — it’s never outdated. This allows him to stay on top of outstanding receivables so that we can better predict cash flows and provide guidance to our branch managers on which customers they need to be talking with.
We’ve also enjoyed significant time savings on our payroll reporting. Reporting on data from multiple systems to the payroll team used to take one of our people an entire workday each week. Now, it’s completely automatic. The time savings and accuracy of all these things helps increase the trust levels in data across the company and has freed everyone up to work on more impactful projects.
Using data to prioritize business needs
With Fivetran, BigQuery and Looker we can prioritize business needs and related workflows. For example, we track our Zendesk support tickets in Looker so we can easily see what’s open, urgent, high priority, pending, and closed. We can also view our open tickets by tag and then identify trends. Similarly, our accounts receivable team can instantly see total outstanding amounts and bills owing. Branch managers can see that information by customer and prioritize who they need to follow up with. Being able to visualize the information intuitively lets us understand what’s important, and act on it quickly.
Looking to the future with telemetry data
I’m excited to add telemetry data to Looker in the near future. Some of our heavy construction equipment and trucks have GPS attached to them so that we can track mileage and location. We’re just beginning to explore use cases to calculate efficiencies in the field with that data. We know we’ll gain a competitive edge by understanding how we can move faster and smarter and by delivering a better customer experience. And, if we can continue to use data more effectively, the efficiencies and time-savings we realize should fully offset the cost of the tools.
A data-driven strategy for future growth
Since modernizing our data stack with Fivetran, BigQuery, and Looker, we’ve reduced the hours spent on manual activities and freed up time to focus on what the data means for our business. My team has been largely freed from spending time providing manual reports from so many disparate systems. We can now focus on strategic initiatives that are going to help us continue growing and serving our customers.
For example, we’d like to add embedded analytics with an external-facing customer dashboard that shows each customer the top metrics they want to know regarding the status of their jobs. We can see that data internally, but it’s powerful to provide our customers with a secured website where they can see milestones, timelines, and financial information about the project, and explore that real time data without having to email us.
Moving from a static reporting and spreadsheet world to a live data environment has been a big culture shift at Emery Sapp & Sons. Everyone understands that they aren’t just saving time and doing less manual work — they are now able to focus on more high-value and interesting work that benefits the company and our customers. They’re getting used to asking and answering their own questions…and everyone trusts the answers.