Interface Metaphors
This week I worked on a high-priority feature. It’s a tool that captures information and generates a long, formal document for security compliance. It sounds simple, but the interaction model sparked a surprisingly deep debate.
The document in question is a known requirement artifact in security frameworks, sometimes reaching 300+ pages, dense with repeated sections and compliance-ese language. Very intimidating. Very complex.
The goal of the tool is to make creating this document fast and painless. Users arrive hoping to get in and out quickly with something auditor-ready.
I spent most of the week trying to talk some stakeholders out of a proposed concept that mirrored the report structure in a block-editor-style form builder. From their perspective, this solution was clear—if the end result is a document, why not use a document editor to build it?
Short answer: it’s not that simple.
Competing metaphors
The concept of interface metaphors was new territory for my stakeholders. They understood the idea in theory, but not its deeper significance in product design.
Interface metaphors help users understand how to interact with a digital system by using familiar, real-world concepts. The desktop metaphor (folders, trash can), e-commerce interfaces (wishlists and shopping carts), AI assistants (chat windows). They leverage the users’ prior knowledge and shape expectations.
In this project, the interface metaphor my stakeholders wanted was a block editor, like Notion or Medium. Sure, it’s great—it allows customization and flexibility. But the end result isn’t a web page or a long-form blog post, it’s a document—with an index, page breaks, headers, and footers.
The first challenge with this approach is that the metaphors were too close. A block editor walks and talks like a document editor, so users will expect rich formatting, layout control, styling, preview modes, and full customization. If I can move blocks around, why can’t I change the page layout? Where are the headers? Can I add a logo or control pagination? How do I preview?
The second challenge was persona mismatch. Our primary compliance-ASAP user isn’t hoping to tailor every paragraph. Good for power users, not for most people. Customizable, but high cognitive weight. The starting point is great, but hardly less intimidating than just doing it in Google Docs.
The last challenge was feasibility. We’re not building a WYSIWYG editing platform. Matching the bar set by tools like Notion or Medium would take months of design and engineering. And even if we did it well, we’d be signing up for a long tail of expectations—tables, embeds, formatting, styles, real-time preview. It’s massive.
The power of abstraction
The concept my team and I sold this week pulls away from the document metaphor entirely. It reframes the experience around structured data capture, abstracting the required information from the final document. This has a few benefits:
- It simplifies the experience by reducing perceived effort. Instead of navigating the full report, users manage information by entity type, leveraging the mental models they already have about their organization and systems. A single data capture can apply to multiple places in the final document.
- It allows us to offer new ways to enter and manage information by pulling away from the interface metaphor of a doc editor. Tables? Fine. Modals, workflows, complex tasks, dashboards? Great. Can’t do these things in a WYSIWYG doc editor.
- It establishes the builder as the single source of truth. The exported document is just a snapshot—something to send to the auditor when needed. It doesn’t need to be beautiful, just accurate.
- It frames the information management as a living, ongoing part of compliance rather than a one-time document creator. This ensures users stay in the system longer, where they can leverage automation, monitoring, and milestones on an ongoing basis. Better for users, better for retention.
ChatGPT was my therapist this week and did some heavy emotional lifting while I argued the validity of my proposal, asking how to frame certain concepts to stakeholders and argue my case. Given the sensitive and high-priority nature of this feature, user research access was very limited. Role-playing and concept validation with an LLM is an acceptable near-term replacement.
I’m proud of where we landed. My PM and engineering lead are aligned and committed to the direction. Tech scoping kicks off Monday, and I’m excited to start building.
Interface metaphors—not just design theory, real leverage.
