ADP Enables Dynamic Benchmarking of Human Capital Management Metrics with Snowflake

ADP provides products, services and experiences that simplify work for more than 1 million clients in 140 countries. Large and small organizations across virtually every industry rely on ADP’s cloud-based human capital management (HCM) solutions to streamline HR, payroll, time, tax and benefits administration. 

Self-service HCM analytics help ADP’s clients understand workforce trends and benchmark their metrics against aggregated, anonymized data from over 30 million employee records. Powering dynamic benchmarking workloads with Snowflake’s Data Cloud enables a seamless user experience while ensuring accuracy, consistency and performance.

ADP’s previous benchmarking architecture required large amounts of precomputing that was time consuming, cost prohibitive, and hard to scale. Precalculating HCM benchmarks in batches led to data freshness concerns and made it difficult for ADP to deliver flexible, frictionless user experiences. “We previously couldn’t offer that flexibility because we had dozens of cubes that would need to be preprocessed,” says Brent Weiss, Senior Director of Product Management, DataCloud Analytics, at ADP. 

Reprocessing data in response to bug fixes, data field changes and evolving business requirements increased cost and complexity at ADP.

Seeking to elevate the customer experience through dynamic benchmarking, ADP turned to Snowflake’s Data Cloud. Snowflake’s elastic performance engine aligned with ADP’s need to maintain subsecond query performance and enable on-the-fly computation. Snowflake’s intelligent infrastructure required less maintenance and oversight, freeing up ADP’s technical staff to focus on higher-impact work instead of synchronizing and regenerating data. 

Moving ADP’s benchmarking logic to use dynamically generated queries with Snowflake resulted in a better experience for users. “Instead of spending a week precomputing data that someone may never ask for, now we can just do it on demand,” says Justo Pastor, Senior Director of Application Development, DataCloud Analytics, at ADP. Snowflake’s multi-cluster shared data architecture made it easier to leverage caching and “warm up” popular queries. According to Justo, “About 90% of the queries we run are subsecond.”

Collaborating with Snowflake’s engineering and product teams helped ADP achieve a “deterministic” approximation function that yielded the desired outcome for the product. “We hit consistency, accuracy and performance just by making a particular change in an approximation function—and the results were profound,” says John Thorpe, Senior Sales Engineer-Financial Services at Snowflake. 

ADP’s reimagined benchmarking engine is transforming the company’s approach to front-end design. “We’re combining NLP with Snowflake to offer a voice-driven interface for users to basically benchmark anything,” says Weiss. 

ADPs next-generation user interface and “conversational” experience will make it even easier for clients to analyze and benchmark against more than 100 data fields and dimensions. According to Weiss, “That’s going to offer a much more powerful experience that leads to client business impact —and improving our clients’ businesses is what this is all about.”

Learn more about the Snowflake Data Cloud and how we measure real-world improvements in performance experienced by customers over time.