Archer Meat Snacks wanted to be seen during a moment when most advertising wasn't even in the room.
On April 7, 2026, GumGum announced results from a campaign that placed full-screen advertisements inside a pause — not before content resumed, not as an intermission, but specifically when a viewer had already decided to stop. The brand achieved +11.5 percentage points in aided brand awareness, +10 points in purchase intent, and +5 points in brand favorability, according to PRWeb. These are strong numbers. They are also numbers that require significant asterisks.
GumGum introduced Pause Ads in February 2026 through a blog post titled "Own the Break: Introducing Pause Ads Built for Viewer Mindset." The product lives inside streaming content on CTV platforms. When a viewer presses pause — on purpose, voluntarily, to step away, to check something, to exist in the break they've created — a full-screen ad appears. It occupies the entire screen. It doesn't interrupt anything, because the narrative was already suspended. It just arrives.
GumGum calls itself "The Mindset Company™" and has operated since 2008 across 19 markets. Its infrastructure is called Mindset Graph™ — an AI engine that processes real-time contextual signals, creative signals, environmental signals, and historical signals to determine what, when, and how to serve.
The metrics are real. Magna and DirecTV Research found in 2025 that 91% of streaming viewers pause regularly, and 81% do so specifically to avoid missing content. TripleLift reported in 2024 that 100% of viewers notice pause ads, with an average dwell time of 24 seconds — dramatically higher than typical in-stream completion rates. DirecTV internal data showed 34% higher unaided ad recall compared to traditional addressable advertising. Wunderkind found 79% lower cost per action versus other CTV formats.
These numbers are real. The asterisks are real too.
Cognitive science has a precise account of what happens when someone suspends a task. The Memory-for-Goals model, developed by Altmann and Trafton in 2002, describes how suspended goals maintain activation during interruptions. When the task resumes, there is a measurable "resumption lag" — the cognitive cost of returning to full engagement. Goals can be inhibited during pauses. The brain has already begun to decompress.
This matters because advertising research has historically treated the pause as a disengagement signal. If the viewer isn't watching, the thinking went, they aren't available. But pause research suggests the opposite: the viewer has created the pause deliberately. They are not between content and distraction. They are between content and themselves.
Kupor and Trafton at Stanford Graduate School of Business published a paper in 2015 in the Journal of Consumer Research titled "Persuasion, Interrupted." Their finding: momentary interruptions can actually increase persuasion compared to uninterrupted messages, under specific conditions. The mechanism involves heightened attention upon resumption — but also, critically, the state of cognitive decompression that precedes resumption. The mind is open. The narrative is not competing for resources.
The pause is not disengagement. It is a self-initiated break that signals high content engagement.
Here is the asterisk that the industry does not prominently display.
The viewers who pause are not a random sample. They are, by definition, the viewers who were paying enough attention to the content to care about not missing it — and who then chose to step away from it anyway. They are, consequently, among the most engaged viewers in the audience. They are also, likely, more receptive to advertising in general.
Research from Hoban and Arora published in 2017 demonstrated that ignoring selection effects in advertising overstates response by more than fivefold. Browsing intensity accounts for a third of this bias; targeting accounts for the remaining two-thirds. Users who receive targeting are inherently more receptive — they were already oriented toward engagement. The same logic applies to pause viewers: the act of pausing selects for viewers with higher baseline receptivity.
This creates a fundamental measurement problem. When Archer achieved +10 points in purchase intent, how much of that lift came from the ad format, and how much came from the fact that the ad was shown to the subset of viewers most predisposed to respond to it?
Krafft, Arden, and Verhoef addressed this directly in a 2017 paper in the Journal of Interactive Marketing on permission-based advertising. They found that the strongest drivers of granting advertising permission are perceived personal relevance and entertainment value — not the ad itself, but the viewer's pre-existing orientation toward the brand. Permission-granters self-select. The measured effectiveness of permission-based formats therefore systematically overestimates the effect on the general population.
The neuroscience of voluntary versus involuntary attention makes this distinction precise.
Landau and colleagues demonstrated in 2007, using EEG, that voluntary and involuntary attention operate via different neural mechanisms. Voluntary attention — top-down, goal-directed, involving prefrontal cortex and beta-band oscillations — pre-allocates processing resources. Involuntary attention — bottom-up, stimulus-driven, captured by novelty, intensity, movement — competes for neural representation but does not guarantee engagement. When a viewer chooses to pause, they are deploying voluntary attention. They have created the moment. They are not being captured by an external stimulus.
This distinction is foundational to how advertising works. Captured attention and directed attention produce different cognitive outcomes. Captured attention gets you noticed. Directed attention gets you processed.
In voluntary attention states, the Advertising Research Foundation's attention validation work found, consumers can move from brief exposure that leaves little memory trace to intense focus with cognitive and emotional engagement that leads to enduring recall. The difference is entirely in the attentional mode.
Pause Ads occupy a voluntary attention state. The question is whether the voluntary attention is directed at the ad or simply available in the room while the ad appears.
CTV advertising has a measurement problem that Pause Ads inherit and, in some ways, amplify.
The standard currency — Nielsen, Comscore, VideoAmp, iSpot.tv — was designed around broadcast and cable. It counts impressions, estimates reach, and produces ratings that reflect population-level exposure. CTV introduces the ability to serve addressable ads to identified households, but the cross-device attribution problem — connecting a CTV ad exposure to a subsequent mobile or desktop action — remains unsolved in any standardized way. There is no universal identifier across smart TVs, consoles, and streaming sticks. Probabilistic matching fills gaps with error.
For Pause Ads specifically, the measurement challenge is three-layered.
First: The format is not standardized. Pause ads on Roku operate differently from Pause ads on Disney+. Custom specs make programmatic scaling difficult and comparison across platforms unreliable.
Second: Most performance reporting relies on surface-level metrics. Impressions served. QR codes rendered. Completion rates. These are proxy metrics for business outcomes, not business outcomes themselves.
Third: The self-selection bias described above — the most engaged viewers are most likely to pause — means that any lift measured among pause viewers overstates what a campaign would achieve if served to a random population of equivalent size.
TripleLift reported that Pause Ads drove a 33% recall increase, 11x brand consideration lift, and 76% purchase increase when added to campaigns. The Zales case study achieved a 276% lift in QR code scans and directly attributable sales. These numbers are directional. They are not the incrementality truth that a brand deciding on a media plan needs.
Nobody is accounting for what would have happened to the control group. Not really.
The viewer has pressed pause. They are sitting in a room with a screen. They have not left — the room is still there, the screen is still on, the content is still loaded in memory. But they have chosen to stop. They are between the narrative and the return. They are in the gap they themselves opened.
That gap is what Pause Ads were built to fill.
The format's advocates describe it as the first truly permission-based commercial moment — not in the opt-in data sense, but in the attentional sense. The viewer created the space. The ad enters a space that was already offered.
This is a real and meaningful distinction. The interruption model of advertising — pre-roll, mid-roll, display — operates by inserting content into spaces the viewer did not request. The permission model — Pause Ads, maybe some native formats — operates by arriving in a space the viewer created.
But the measurement problem remains unresolved. The lift numbers are real. The causal story is not yet proven.
What Pause Ads represent is a threshold — a genuine threshold in how advertising relates to viewer intention. Whether that threshold leads somewhere better depends on whether the industry can build the measurement to match the moment.