DATABASE GOVERNANCE | APPROVAL WORKFLOWS | ROLE-BASED UX
DATABASE GOVERNANCE | APPROVAL WORKFLOWS | ROLE-BASED UX
Unrestricted database access allowed any user to instantly alter live content across a customer’s digital signage network. As Principal Product Designer, I designed a row-level status architecture that embedded data governance directly into the spreadsheet interface, replacing manual off-platform communication methods with in-app approvals across thousands of endpoints.
If you have limited time, just read the → arrows for a 60 second summary of my strategy & product impact.
PROJECT OVERVIEW
The platform's online data grid was a highly flexible spreadsheet tool, but its biggest strength was also its greatest vulnerability: it had no guard rails. Anyone with access could edit a cell and instantly change data on public facing digital signs. Enterprise clients forced their teams into an off-platform manual review process that included cluttered text messages, screenshots, and external email threads just to verify changes. The product lacked the structural governance needed to let teams collaborate safely at scale.
We introduced a row-level status and approval workflow built directly into the existing grid layout. Instead of forcing users to utilize off-platform tools or separate administrative hubs, data changes are now staged, flagged, and locked line-by-line. Administrators can review, comment on, and approve updates natively inside the data grid. The messy communication loop has been solved entirely in-platform.
I owned the user experience and interface architecture for this governance layer. The scope required integrating strict user permissions and multi-state workflows while preserving the rapid, high-density editing performance users expected from a spreadsheet layout. Additionally, the UX had to work within rigid backend database constraints that prevented granular change histories or live visual diffs.
Pulled 100% of the review cycle out of fragmented email and chat threads, centralizing all change histories and rejection context directly within the platform.
Eliminated critical public-facing content errors across diverse signage networks by enabling administrative safeguards on pending content updates.
Maintained the rapid, high-density data entry speed of the spreadsheet interface while seamlessly introducing enterprise-tier compliance workflows.
USER FRICTION
Before this feature, managing data inside the platform's online data grid felt like walking a tightrope without a net. To understand how our users coped with a lack of governance, I conducted interviews with enterprise content administrators at major resort networks and corporate campuses.
The feedback was unanimous: they were fixated on the risk of their users publishing un-approved content to public endpoints. Because any typo or incorrect image would instantly go live on a massive public display, administrators resorted to using highly inefficient workarounds to protect their brand integrity:
Teams maintained separate "staging" spreadsheets in Excel or Google Docs where regional coordinators would input data first.
Administrators required users to email or Slack them screenshots of their proposed changes before copy-pasting them into FWI Tables.
In many cases, organizations completely locked out local users who desperately needed to update their own signs, forcing a handful of central administrators to manually handle hundreds of data entries a week.
WORKFLOW MECHANICS IN THE GRID ARCHITECTURE
Our users loved the online data grid as a product because it gave them the high-density editing speed of a standard spreadsheet, but had the benefit of being tied into their digital signage ecosystem. If we forced them to go through a more burdensome process like a multi-step wizard or a wholly separate workflow to approve a single line item, we would diminish the core value of the tool.
Partnering with engineering, I structured the approval logic so that the entire row updates as a single package. When a user modifies a cell within a row, the entire line item shifts into a Pending state. To maintain readability in a high-density grid, I avoided heavy alerts. Instead, the UI gently dims the text and applies a subtle background tint across the row. This signals to the contributor that their work is saved but not live, while simultaneously locking the row from conflicting edits until an administrator takes action.
To make administrative controls obvious without suffocating the existing data fields, I introduced a dedicated Workflow column anchored to the far left of the grid layout.
Because wide enterprise tables introduce horizontal scrolling friction, I anchored the workflow controls to a sticky column on the far left, ensuring statuses and actions remained permanently visible, no matter how wide the data grid grew.
While inline single-row actions worked perfectly for quick edits, enterprise users often face waves of updates at a single time. To prevent the interface from becoming a clicking chore, I ensured the solution accounted for multi-row interaction and management. Admins and submitting users alike can select a batch of rows, which then surfaces global actions to submit, approve, or reject in bulk.
THE FEEDBACK LOOP
The biggest workflow bottleneck we identified in research was the "Why." When an administrator rejected a change, they had to go to email or Slack to tell the contributor what to fix.
To solve this, I designed a context-aware inline popover. Clicking "Reject" opens a tight, focused dialogue box attached directly to the specific row. The administrator can type a concise note, which pins the comment directly to that record.
When the contributor logs back in, they see a rejection flag on their row, giving them access to see the exact feedback, make the edit, and resubmit. 100% of the workflow communication is kept inside the platform.
RETROSPECTIVE
While initial discovery leaned toward a separate approval module, I integrated the governance layer directly into the existing grid view. This optimized current user patterns and reduced development cycles while minimizing onboarding friction.
Don't break the user's mental model. When users spend hours a day in a spreadsheet, they develop extreme muscle memory for data density and speed. The moment you pull them out of that layout to complete an administrative task, you introduce cognitive friction. By keeping the entire approval ecosystem inline, we proved that enterprise needs don't have to break daily operational flow.