Progressive Disclosure Is Gone. Engagement Metrics Killed It.

The print dialog in Microsoft Word circa 1998 had 47 settings. Most users needed two of them. The design principle that solved this problem is nearly thirty years old, well-documented, and no longer applied anywhere that matters.

Progressive disclosure is the UX principle of showing essential features first and deferring advanced options until the user actually needs them. Bruce Tognazzini and Jakob Nielsen formalized it at the Nielsen Norman Group — their 2006 article is still the canonical reference. The principle reduces cognitive load, speeds task completion, and cuts the error rate from accidentally triggering features users didn't know existed. What neither Tognazzini nor Nielsen modeled in 2006 was what happens when the optimization target changes from "task completion" to "time spent in app." On the second metric, progressive disclosure is a liability.

What Progressive Disclosure Did

The principle is straightforward: users don't need to see everything at once. Show the 20% of features that handle 80% of tasks. Put the rest behind a tap, a toggle, an "advanced settings" flow. The user encounters a simpler surface. They learn at their own pace. Errors from accidentally activating unknown features drop.

The canonical illustration is the printer dialog. Print a document: you need paper size, page range, number of copies. You don't need duplex binding configuration, watermark settings, or ICC color profiles. Progressive disclosure surfaces the first group immediately and collapses the second behind "More Options." Time to print drops. User errors drop. Satisfaction increases.

This worked because it was designed for goal-oriented tasks with a clear completion state: print the document, send the email, book the flight. There is a "done" condition, and the interface's job is to help the user reach it efficiently. The complexity existed but was deferred until it was needed.

Where the Metric Changed

Instagram launched in 2010 with a simple feed, a follower button, and three interaction options: like, comment, and share. By 2023 the main screen presented Stories, Reels, a Shopping tab, a notification badge, a DM count with unread indicator, a persistent five-item navigation bar, and creator badges pinned to posts. The 2023 redesign that was supposed to simplify the interface collapsed some tabs while adding persistent Quick Reply access to the main feed.

None of this is progressively disclosed. It's all visible from the first frame. The logic is not "show users what they need." It's "expose users to as many features as possible so they find something to engage with, and stay longer while they're looking."

Instagram's Q3 2023 earnings call made the tradeoff explicit. Time spent on the platform increased 24% year-over-year. Net Promoter Score — tracking whether users would recommend the product to others — dropped to its lowest recorded level in the same period, before recovering partially in Q4. The company reported both metrics. It continued optimizing for the first.

The TikTok Inversion

TikTok's video feed is the complete architectural inversion of progressive disclosure. Every video in the feed simultaneously displays: a like button, comment button, share button, creator gift button, follow button, profile link, audio title, duet option, and stitch option. On a standard 6.1-inch phone this fits as a stacked icon column on the right edge of the screen. A new user encountering this interface has no clear indication of what half the icons do.

The design relies on habituation instead of guidance. Users learn through accident — they hit "Stitch" when aiming for "Comment," discover what it does, occasionally use it intentionally. The interface doesn't teach. It exposes. Discovery happens through error or through long enough scroll depth that the feature appears in context.

This produces the engagement numbers TikTok needs. Average session length on TikTok in 2024 was 95 minutes. It also produces what UX researchers call feature blindness — users systematically missing features they would use if encountered contextually, because those features were shown too early and with no task frame to attach them to.

Related: Nobody Knows How to Design for AI Agents Yet covers the distinct challenge of designing for systems where the full capability set genuinely cannot be previewed upfront — a problem that requires progressive disclosure to be reinvented, not abandoned.

What Got Lost in the Switch

The task completion cost is real and shows up in app review language. Search App Store reviews for any major social platform and find: "I can't figure out how to find old posts," "the settings to manage who can see this are impossible to find," "I can't figure out how to edit something I already submitted." These are users who would have found these features under a progressively disclosed interface — contextual menus, help flows, an "advanced" reveal path. They're invisible in the all-features-visible design because they're surrounded by options they didn't ask for.

The cognitive load literature has established for decades that increasing the number of visible options increases decision latency and error rate for users performing goal-oriented tasks, even when the added options are genuinely useful. Feature exposure doesn't translate to feature discovery in task contexts. It translates to confusion, longer paths to completion, and a higher abandonment rate on tasks that the user started with clear intent.

The populations most affected are the same ones accessibility frameworks try to protect: users with lower working memory capacity, unfamiliar platform users, and users in cognitive fatigue — all of whom benefit most from seeing less at once and are harmed most by maximum feature exposure at load time.

Progressive Disclosure Isn't Dead, Just Displaced

Enterprise software kept it. Figma's inspector panel surfaces the properties relevant to whatever you've selected — switch from a text layer to a frame and the panel changes. Notion's slash command menu is progressive disclosure by design: type / to surface the command list, narrow it by typing more characters, and the feature set reveals itself proportionally to your specificity. Linear shows task status and then expands to full detail on click, keeping the list view scannable.

These products optimize for task completion. Their users need to ship code, write documents, close bugs. The product cannot afford to trade task completion for dwell time — users would notice the friction immediately and find a replacement. The economic incentive aligns with the design principle.

The principle works where it always worked. The question worth asking of any interface is whether it was designed for you to accomplish something or for you to stay. The answer shows up in whether the interface reveals what you need when you need it, or puts everything in front of you at once and calls the resulting confusion discovery.

Photo: AS Photography / Pexels