The Scroll Is Dead: How Vertical Feed Design Is Being Replaced

For about fifteen years, the scroll was the dominant metaphor of the internet. You opened something, you scrolled down. Down and down and down, with no bottom in sight. It was so ubiquitous that most designers stopped questioning it. It was just how digital products worked. That assumption is now breaking down in ways that are visible, documented, and accelerating.

Who Invented the Scroll and Why

Infinite scroll was filed as a patent by the Windows Live Image Search team at Microsoft in February 2006 and shipped the following month. The logic behind it was genuinely good: for a grid of image results, pagination forced users to stop, click, wait, and reorient every time they reached the bottom of a page. Continuous loading removed that friction and kept users moving through content.

The pattern made sense in that specific context. It made sense when Facebook adopted it. It made sense when Twitter built its feed on it. For content that was naturally continuous and discovery-oriented, where users had no specific destination and were comfortable drifting, infinite scroll was a reasonable design choice.

The problem is what happened next. Infinite scroll stopped being a considered design decision and became a default. If you were building a feed, a product list, a news page, or almost anything with multiple items, you used infinite scroll without much debate. It was what the internet looked like. The psychological mechanics behind why it worked so well at keeping people on screen were understood by designers and product teams, but the question of whether they should be exploited at every opportunity was rarely asked.

The Backlash Built Slowly, Then All at Once

The American Psychological Association was warning about infinite scroll and its particular risk to younger users well before the mainstream conversation caught up. Research published through the CHI conference in 2025 formalized what had been accumulating for years: infinite scroll is not a neutral design pattern. It is one of the primary mechanisms through which people end up scrolling mindlessly on social media, not because they want to, but because the design removes the natural stopping points that would otherwise let them decide they have had enough.

The neuroscience behind this is not complicated. Each new item that loads as you scroll produces a small dopamine response. The brain gets into a rhythm where re-engagement with each new item requires minimal mental effort. Users often recall the content they consumed but not where they saw it, because the lack of boundaries between pieces of content means positional memory never forms. You cannot remember where something was because there was no meaningful sense of where anything was.

By 2024 and 2025, the institutional response had become concrete. Instagram introduced daily limits and break reminders. YouTube added settings to restrict time on the platform. The UK's Online Safety Act introduced provisions aimed at addictive features in platforms serving children. Several EU member states moved to regulate infinite scroll as part of broader digital wellbeing legislation. The design pattern that had defined the social web for a decade was formally designated as risky by governments and public health bodies.

Three Products That Broke the Pattern First

The real signal that infinite scroll was losing its dominance came not from regulators but from products. Three specific platforms changed the model in ways that proved themselves commercially, and in doing so gave the industry a different set of reference points to work from.

TikTok and the vertical swipe

TikTok did not use infinite scroll. It used discrete vertical swipe. Each video was a complete unit. Swiping up moved you to the next one. The content did not flow together into a continuous stream. Each piece had a beginning and an end, and the gesture to move on was an explicit, conscious action.

This matters more than it might seem. The swipe is an active choice. The scroll is a passive continuation. TikTok's interaction model put a micro-decision between each piece of content and the next. Paradoxically, this did not reduce engagement. It increased it. Users felt a greater sense of agency, which made them more willing to engage rather than less. YouTube Shorts and Instagram Reels followed the same model for the same reasons.

The vertical swipe is not just aesthetically different from the scroll. It is cognitively different. It reintroduces the boundary that infinite scroll was specifically designed to eliminate.

Perplexity and the single-answer terminal

Perplexity AI built its product around a fundamentally different relationship with scrolling. You ask a question. You get an answer. The answer occupies a finite amount of space. When it ends, it ends. There is no feed below it. There is no algorithmically surfaced content waiting to pull you further down the page.

This is not a scroll alternative. It is a scroll rejection. The interface model says: you came here with a question. Here is the answer. You are done. This sounds radical until you realize that it describes how most productive software used to work before social feed design colonized the entire internet. A search result page, a document, a dashboard: all of these are finite. The assumption that content should extend indefinitely was imported from social media and applied to contexts where it never made any sense.

Perplexity's commercial success with this model has been significant. It is not succeeding despite rejecting infinite scroll. It is succeeding partly because of it. Users who are trying to accomplish something do not want to be held in a feed. They want the answer and then they want to leave.

Arc, Raycast, and Linear: keyboard-first, no scroll needed

A different set of products broke the scroll paradigm from a completely different direction. Arc browser, Raycast, Linear, and Superhuman are all products where the primary interaction model is keyboard-based navigation rather than scroll-based browsing. You move through these products by typing, not by scrolling.

This is a revival of a much older interaction model, the command-line approach, applied to modern product design. In Raycast, you invoke a command prompt and type what you want to do. The entire product is navigated this way. Scrolling through a list of options is a fallback, not the primary mechanism. In Linear, keyboard shortcuts are so central to the experience that the product teaches them actively rather than hiding them in a settings menu.

These products have built intensely loyal user bases, particularly among developers and power users. Their rejection of scroll-as-navigation is not a quirk. It is a design statement about what kind of interaction these products are optimized for: fast, intentional, purposeful.

What Is Actually Replacing the Scroll

The honest answer is that infinite scroll is not being replaced by a single successor. It is being replaced by a set of patterns, each suited to a different context and a different relationship between the user and the content.

None of these patterns is universally better than infinite scroll. The point is that each one is more appropriate than infinite scroll in the context it is designed for. Infinite scroll was applied everywhere regardless of context. The shift happening right now is toward more deliberate pattern selection based on what the user is actually trying to do.

Where Infinite Scroll Still Makes Sense

Being clear about this matters. Infinite scroll is not a bad pattern. It is a pattern that was applied far beyond the contexts where it performs well.

It remains genuinely suited to entertainment-first content where the user has no specific goal and is comfortable being pulled through a continuous stream. A product like Pinterest, where discovery is the entire point and there is no meaningful concept of where you are in the content, works well with infinite scroll. A social feed where the primary purpose is passive consumption and the user experience is genuinely improved by removing friction is a reasonable context for it.

The Baymard Institute has been consistent on this point. Their research on e-commerce UX found that infinite scroll performs poorly for product lists precisely because users need positional memory. If you scroll past a product and want to find it again, an infinite scroll interface gives you no way to locate where you were. Pagination, or at minimum a robust scroll position memory system, serves e-commerce users better. For entertainment feeds, the calculus is different.

The question that most teams are still not asking carefully enough is: which context does my product actually fit? The answer to that question should drive the navigation pattern choice, not the default assumption that scroll is how digital products work.

What This Means for the Products You Are Building

The practical implications of this shift are real and worth working through.

Your users are arriving with different expectations than they had five years ago

Users who spend time in Perplexity, Arc, or Linear bring a different set of expectations to every other digital product they use. They have experienced interfaces that respect their time and give them a finite, purposeful interaction. When they encounter an interface that holds them in a feed, the contrast is more jarring than it used to be. Expectations are being reset by the best products in each category, and the tolerance for manipulative scroll design is lower than it was.

The context of use should determine the navigation model

If your users come to your product to accomplish something specific, a feed model is probably working against you. Task-oriented users need to be able to find what they came for, complete what they came to do, and leave. An interface that keeps them scrolling past additional content on the way to their goal is not serving them. It is serving a time-on-site metric that does not translate to satisfaction or return visits.

If your users come to your product to explore or discover, without a specific destination, scroll-based navigation may still be appropriate. But even here, the discrete swipe model offers advantages worth considering. It reintroduces micro-decisions. It creates natural stopping points. It makes the experience feel more like a product and less like a mechanism.

Keyboard navigation is no longer optional for power users

The success of Linear, Raycast, and Arc has demonstrated that a meaningful segment of the user population strongly prefers keyboard-first navigation in tools they use heavily. These users are often among the most valuable in a product's customer base: high engagement, high tenure, high word-of-mouth influence. Building robust keyboard navigation is not an accessibility bonus feature anymore. For products with power users, it is a retention feature.

Positional memory is a feature, not a given

If your product uses infinite scroll, your users are losing their place constantly. They go back from a product page and the list jumps to the top. They find something interesting and cannot locate it again when they return. This is not a technical limitation. It is a design choice, and it is one that can be addressed without abandoning scroll-based navigation entirely. Scroll position persistence, session state memory, and explicit anchor links all restore the positional memory that infinite scroll removes.

A Practical Audit for Your Current Navigation Model

If you want to understand whether your current scroll design is serving your users or working against them, these are the questions worth asking.

• What do users come to your product to do? If the answer is a specific task, map the path from arrival to completion. Count the scroll depth required to reach the goal. Every scroll that is not moving the user toward their goal is friction.
• Can users find things again after they have seen them? Test this directly. Ask a user to find something they saw ten minutes ago. If they cannot, positional memory is broken. This is a direct cost to the user and to your product.
• Where do users stop scrolling? Heatmaps show scroll depth. If most users stop well before the bottom of a page, you have content that is not being seen by the majority of visitors. Pagination or a load-more pattern at that depth would serve them better than an infinite continuation they are not reaching.
• Are users actively choosing to continue, or just continuing? There is a meaningful difference between a user who scrolls because they want more and one who scrolls because there is no reason to stop. Session recordings show this. Users who pause, orient, and then scroll with apparent purpose are engaged. Users who scroll continuously at a constant speed with no pauses are running on autopilot.
• What do your most loyal users say about navigation? Long-tenured users who use your product heavily have developed strong opinions about its navigation model, usually through accumulated frustration with the parts that do not work. Talking to them is faster and cheaper than running a study.

The Scroll Was a Default. Now It Is a Choice.

That is the shift. Not that scroll is wrong. Not that every product should replace it with a command palette or a vertical swipe. But that scroll is now a choice that needs to be made deliberately, justified by the context, and designed to serve users rather than to hold them.

The products leading this shift are not doing it out of altruism. They are doing it because users reward them for it. The products that have moved away from manipulative scroll mechanics in favor of purposeful, bounded, task-respecting interfaces are building more loyal user bases than the ones still competing on time-on-screen.

That is the real story here. Not that the scroll is dead. It is that designing for what users actually want to do has started to outperform designing for how long you can keep them on the page. The scroll was a tool for the second goal. The products winning right now are the ones optimized for the first.