Building Compass

Replacing a fragmented legacy platform with a purpose-built system for school-facing workflows, data validation, reporting, and accountability.

Building Compass

In July 2025, DC Public Charter School Board faced a crisis of confidence among the public charter schools it oversees. Repeated data errors and fragmented workflows had damaged trust with our local education agencies. The agency depended on commercial and low-code tools that supported essential operations but made those operations difficult to standardize and improve.

I proposed building a web application that DC PCSB would own and control. As Director of Data and product owner, I defined the product vision, secured support for the investment, formed and staffed a Scrum team across product, design, and engineering, and led Compass from concept to a live minimum viable product in 10 months.

The launch established that the agency could build and operate its own platform. The larger undertaking is to replace fragmented processes with reusable capabilities for collecting information, validating data, coordinating review, and delivering reports. Compass remains under active development, with migration proceeding one workflow at a time.

Why the Hub needed to change

DC PCSB’s oversight responsibilities require sustained coordination with schools. A single process can involve several school staff members, multiple agency teams, supporting documents, validation rules, corrections, and a final decision. Deadlines and responsibilities must remain clear throughout, and the information collected often becomes the basis for subsequent reporting or accountability work.

The agency’s existing platform, the Hub, handled many of these processes through QuickBase. Over time, however, similar workflows developed in different ways. Data models were difficult to standardize, permissions and status tracking became complicated, and users had to learn variations in how individual processes worked. The platform’s technical constraints shaped what the agency could build.

Those constraints were compounded by unclear product ownership. New functionality accumulated without a coherent roadmap or a consistent architecture for external submissions and internal coordination. Replacing QuickBase would give the agency more technical flexibility, but realizing that flexibility required clearer priorities, ownership, and product management.

Building around recurring needs

The central product decision was to build reusable capabilities around the patterns shared by oversight workflows. Assigning a task, collecting structured information, accepting files, validating a submission, requesting corrections, recording a decision, and delivering an output are recurring needs across the agency.

Compass is designed to make those capabilities behave consistently. School staff should be able to recognize how a new task works because they have completed another task on the platform. Internal teams should be able to introduce a process by combining established components, reserving new development for needs those components cannot yet support.

That changes the question asked during product development. Instead of treating every request as a separate workflow to construct, the team examines which capabilities already exist, which aspects of the process need redesign, and whether a proposed variation is necessary. Standardization becomes a product responsibility rather than something developers must reconstruct after the fact.

My role has been to connect those decisions to the agency’s operational needs. I have prioritized competing requests, reviewed workflow and interface designs, coordinated staff and contractors, clarified decision rights, and shaped how Compass should integrate with the agency’s data architecture. Throughout, the team has had to develop the replacement while supporting the current-year work that still depends on the Hub.

Connecting workflows to governed data

Compass is one part of a broader enterprise data architecture. Schools submit or validate information through the application; governed systems maintain the resulting records; the warehouse provides authoritative data; and the Data Mart supports documented access for analysis and reporting. Quarto and other reporting tools can then produce standardized outputs for delivery through Compass.

The intended result is continuity between a submission and the information produced from it. A workflow should not end with an attachment that someone must manually transfer into another system. The submission, validation, stored record, and resulting report should remain traceable parts of the same process.

Compass is also intended to provide a common delivery point for the agency’s reporting portfolio. The Reporting Center is designed to help users find reports appropriate to their roles and permissions, while giving the agency a way to govern access, retire outdated versions, and understand which products people use. This connects the platform’s submission functions to the broader objective of returning useful, governed information to schools and staff.

Learning through production use

The first live school-facing workflow supported LaunchPad configuration validation for the upcoming school year. Schools were asked to confirm information about their accountability configurations before those settings were used in later processes.

This was a deliberately bounded introduction. The task mattered operationally and exercised core capabilities, including authentication, assignment, instructions, validation, status tracking, and internal review. It allowed the team to learn from actual school use without requiring the entire legacy platform to be replaced at once.

Production use also made support part of product development. The team needed to respond to questions, correct problems during the operational window, and examine where users encountered confusion. Repeated questions could point to unclear instructions, an interface problem, an awkward workflow, or a policy ambiguity. Each explanation calls for a different response.

The long-term objective is to reduce the need for support by improving the underlying product. Resolving a ticket completes an immediate obligation; understanding why it occurred can improve every subsequent workflow that uses the same capability.

Managing the transition

The Hub continues to support work that cannot move immediately. Introducing Compass incrementally means that schools and agency staff may need to use both systems during the transition. It also creates pressure to reproduce familiar workflows quickly, including variations that the new platform is supposed to eliminate.

Managing that pressure requires a clear architectural direction alongside practical migration decisions. Each workflow offers an opportunity to determine what should be standardized, redesigned, or retired. Deadlines and capacity may constrain the timing, but they should not quietly become permanent reasons to preserve avoidable complexity.

Sustained product management is essential to that discipline. Compass needs someone accountable for understanding users, evaluating requests, maintaining priorities, and deciding what the platform should not do. Without that capacity, greater technical flexibility could simply allow another accumulation of disconnected features.

The same discipline applies to interface design. School leaders, data staff, agency reviewers, analysts, and support staff have different responsibilities and levels of technical knowledge. The application needs to explain what each person must do, what information is incomplete, who owns the next step, and when the task is finished. Users should not have to infer the process from the structure of the software.

What the work demonstrates

Compass has moved from a proposal to a working application with live school-facing use. That is a meaningful delivery milestone, but it does not yet establish the full value of the replacement. The next test is whether reusable capabilities can support additional workflows reliably while reducing confusion, manual coordination, and dependence on the Hub.

The initial investment is substantial. The financial case anticipates that Compass will become less expensive within several years than the combination of tools and custom work it replaces. Those savings remain an expectation to test as migration proceeds.

For me, the project has made the relationship between data leadership and product leadership concrete. Reliable data depend partly on the processes through which people submit, review, correct, and use information. Improving those processes requires decisions about ownership, policy, architecture, and user experience as well as engineering.

The next phase is to expand Compass while preserving that coherence: migrate additional workflows, strengthen shared components, connect more processes to governed data, and retire legacy functions as replacements become dependable. Completion rates, support demand, adoption, and user feedback will help determine whether the platform is making the work easier.

The objective is a common operating platform through which DC PCSB and public charter schools can coordinate data, validation, accountability, and reporting work. Achieving it will depend on how consistently the agency manages Compass as a product and how much unnecessary work the platform ultimately removes.