The Brand Reset — What dentsu's 2026 Attention Study Means for SMB Operators

Study at a glance

The largest video-effectiveness dataset to date, engineered to link attention to long-term brand equity.

The Brand Reset is an industry-scale single-exposure study measuring how video advertising on ten next-generation platforms — alongside Linear television — contributes to both short- and long-term sales outcomes. Its methodological signature is the integration of Lumen's eye-tracking attention data with Kantar's Meaningful-Different-Salient (MDS) brand equity framework and the BrandZ database, which together permit the projection of a single exposure into a three-year sales-lift estimate. The study was conducted in the United States and the United Kingdom.

40,000Respondents

3,600Creative units tested

10 + TVPlatforms measured

9Industry verticals

Figures from dentsu, The Brand Reset, 2026. Methodology: Kantar Context Lab single-exposure design; attention weighted by Lumen's eye-tracking dataset; sales projections via Kantar's BrandZ-driven simulator.

Why it matters for SMB and mid-market operators

You are structurally exposed to the loop this research describes.

The attention-quality crisis documented in the study is not a problem of giant brands with oversized media plans. It is most acute among advertisers whose budgets are scrutinised quarterly, whose attribution stacks default to last-click, and whose CFOs ask for proof by next week.

dentsu describes a recurring failure pattern in contemporary digital advertising that it names the advertising doom loop — a formulation drawn from the 2025 WARC analysis The Multiplier Effect. The loop's shape is familiar to any operator who has managed a Google Ads or Meta account under quarterly revenue pressure:

The Advertising Doom Loop — after dentsu (2026) and WARC (2025)

Stage 1

Budgets are optimised against measurable short-window metrics — CTR, CPA, 30-day ROAS — because those are the only numbers the finance team will approve.

Stage 2

Creative and media progressively commoditise as the algorithm rewards whichever unit produced yesterday's cheapest click. Playback effects plateau.

Stage 3

Because brand investment can no longer be defended on a quarterly P&L, remaining brand spend is reallocated to "performance formats", and the loop restarts at a lower baseline.

Result: a slow erosion of brand equity that is rarely visible in weekly reporting, but that EY-Parthenon and others have labelled the "silent killer" of small and mid-sized advertisers.

SMB and mid-market operators in retail, automotive, insurance, and home services carry three aggravating conditions that make the loop hit harder: compressed decision cycles that reward short-window metrics, no in-house brand-tracking infrastructure to catch the erosion early, and a mental model of "TV is for brand, digital is for clicks" that rules out the one intervention the research says would break the cycle. The Brand Reset supplies the missing evidence base — and, read carefully, the beginning of a way out.

Five findings, five actions

What the study concluded — and what to change on Monday morning.

Each of the following findings is attributed directly to dentsu's study. The analysis that follows each stat is a Signostic interpretation for the SMB and mid-market context. Citations point to the corresponding passages in the original report.

FINDING 01

Digital video — including short-form — delivers multi-year brand-building effects.

Linear television produces approximately a 4.5% long-term sales lift from a single exposure; short-form digital video produces approximately 2%, but at a materially lower cost per percentage point of lift gained.Source: dentsu, The Brand Reset, 2026 — Actionable Insight 01.

Signostic translation

The dominant mental model in SMB media buying is that brand-building is the exclusive province of broadcast television, and that digital video exists only to harvest near-term demand. The study's single most commercially consequential finding is that this model is empirically wrong. Short-form digital video in the hands of SMB advertisers — a YouTube pre-roll unit, a Meta Reel, a TikTok in-feed spot — is measurably doing long-term brand-building work on a three-year horizon, whether or not the media buyer intended it to. Evaluating that investment solely against a 30-day ROAS figure is therefore not conservative accounting; it is a systematic undervaluation of the asset.

What to change Monday morning

Introduce a second reporting window alongside your 30-day ROAS: a 12-month retrospective view of branded search query volume, direct-site sessions, and unaided recall (where you can afford a simple pre-post survey). Any digital video investment that has moved those secondary metrics is earning back value you have been writing off.

A question worth asking

Of the video spend currently running in your account, what proportion is being judged on a window shorter than the attribution window implied by your category's actual purchase cycle?

FINDING 02

A single exposure can generate 1–5% long-term sales lift alongside a 2–15% short-term lift.

Modelled against Kantar's sales simulator, a single advertising exposure produces a projected 1–5% additional sales lift over the following three years, in addition to a measured 2–15% short-term lift over the following three months.Source: dentsu, The Brand Reset, 2026 — Actionable Insight 02; validated against Binet's commentary on audiovisual memory formation.

Signostic translation

Every qualified impression performs two commercial jobs simultaneously. One is the job your attribution model already rewards: contributing to this week's conversions. The second is the job your attribution model is not built to measure: laying down memory structures that will surface in the category entry points of future buyers. The study's effect sizes are substantial enough that, for most SMB categories with finite customer bases, the cumulative three-year contribution of a year of video spend can meaningfully exceed the three-month lift your dashboard currently records. Media budgets built only around the short window are, in effect, amortising a three-year asset over three months and then writing it off.

What to change Monday morning

Reframe the language of your media plan: replace "cost per impression" with "investment per impression, carried forward." Ask your reporting to estimate the long-run value of this quarter's impression volume at a conservative 1% lift figure, and compare it to the one-quarter CPA you normally report against. For most SMB accounts the delta will change how the next budget conversation goes.

A question worth asking

When was the last time your reporting attempted to estimate the three-year carry-forward value of a quarter's media spend, rather than the quarter's attributable revenue alone?

FINDING 03

Connected TV delivers brand-building effects on par with Linear television.

Connected television platforms deliver a 3.21% long-term sales lift per single exposure, against 4.43% for Linear TV — a near-equivalence that represents a structural change in the media landscape and, according to the authors, dismantles Linear TV's long-presumed supremacy as a brand-building medium.Source: dentsu, The Brand Reset, 2026 — Actionable Insight 03.

Signostic translation

Linear television is economically unavailable to most SMB and mid-market advertisers; even a modest regional broadcast buy typically sits above the monthly media budget of a local retailer, dealership, insurance broker, or home-services contractor. For the past decade, "no television" has therefore meant "no brand-building channel" — a resignation rationalised by the belief that performance digital would carry the load. This finding collapses that rationalisation. Connected-television inventory is purchasable on a programmatic CPM through self-serve DSPs; the minimum commercial spend is small enough to fit a five-figure monthly budget; and the brand-building effect, per the study, is within a percentage point of the linear benchmark. The category that until recently was locked out of the brand-building channel now has a credible path in.

What to change Monday morning

Request a quote from at least one programmatic DSP that offers self-serve CTV (Basis, Viant, StackAdapt, or the YouTube-native CTV buying path). Reserve 5–10% of your current video budget for a three-month CTV test. Measure against the secondary metrics from Finding 1, not against a 30-day ROAS.

A question worth asking

If the economic argument against television has been "we can't afford it," does that argument still hold when the long-term lift per exposure on CTV is within a single percentage point of the linear benchmark?

FINDING 04

Voluntary attention outperforms forced attention, second for second.

Skippable ad formats, which grant the viewer the option to leave, produce weaker effects at one to two seconds of view time — but once viewers elect to continue watching, their brand-building impact surpasses that of non-skippable formats at equivalent attention durations. The authors interpret this as evidence that voluntary, earned attention operates in the same psychological register as earned media.Source: dentsu, The Brand Reset, 2026 — Actionable Insight 04.

Signostic translation

The SMB buyer's default prejudice against skippable inventory is that it is a worse-quality impression than a guaranteed view. The study inverts that prejudice: a skippable impression watched voluntarily is a stronger brand signal, precisely because the viewer chose it. The planning axis that matters is therefore not TV-versus-digital, nor long-form-versus-short-form, but whether the viewer held attention because they had to or because they wanted to. This shifts the centre of gravity in creative production. The first two to three seconds of a skippable unit are no longer the low-risk soft opening they were in the era of the 30-second television spot; they are the entire qualifying round.

What to change Monday morning

Audit your existing video creative inventory against one criterion: in the first two seconds, without sound, would a disinterested viewer have a reason to keep watching? Any creative that fails this test is structurally unable to earn the voluntary attention the study shows is the carrier of brand-building value.

A question worth asking

What fraction of your in-flight video creative is engineered for the opening two seconds rather than a legacy broadcast storyboard that assumes guaranteed viewership?

FINDING 05

Attention matters, but delivers diminishing returns after roughly twenty seconds.

Additional seconds of attention beyond approximately twenty seconds contribute little further long-term lift, consistent with established theories of rehearsal, saturation effects, and memory encoding. The authors explicitly question the commercial case for video units longer than thirty seconds.Source: dentsu, The Brand Reset, 2026 — Actionable Insight 05; Les Binet, advisory commentary.

Signostic translation

The thirty-second television spot remains the unit length of most reusable broadcast assets in SMB marketing archives. If the long-term effect of a spot is substantially complete after twenty seconds of attention, the implication for budget efficiency is direct: every additional second of inventory above that threshold is, at best, a rounding error on brand effect and, at worst, the reason you are buying a more expensive placement than you need. For an SMB operator reusing existing television creative on digital channels, the cheapest optimisation available is often a tighter cut.

What to change Monday morning

Commission a 15-second edit and a 20-second edit of your single best-performing video asset. Run both against the original in a low-spend A/B over four weeks. In our audits, shorter cuts typically survive the comparison on every metric that matters.

A question worth asking

When was the last time you tested a sub-twenty-second cut of your longest-running brand creative, and what did the attention-weighted response look like?

The diagnostic framework

Mapping Kantar's MDS framework onto media your budget can actually buy.

Kantar's Meaningful-Different-Salient framework is the analytic backbone of the Brand Reset. It is not a Signostic construct — the framework predates the study by many years and is proprietary to Kantar, with a published validation record tying strong MDS scores to superior long-run financial performance in the BrandZ database. What the study adds is a per-component attention curve: the amount and type of attention required to move M, D, and S differs by task. Signostic's contribution here is one further layer: translating that task-to-attention mapping into the narrow set of media units that SMB and mid-market budgets can reach.

Task-to-format mapping

Framework: Kantar MDS (from dentsu, The Brand Reset, 2026). SMB application: Signostic.

S · Salience

Be remembered first at the moment of need.

Attention requirement: short, repeated. The study shows salience effects can accrue on attention windows as brief as three seconds.

SMB-accessible formats: Meta Reels, TikTok in-feed, YouTube Shorts, programmatic display on local endemic inventory, Google search with distinctive branded assets.

Priority: attentive reach over attentive time. Budget for frequency, not duration.

M · Meaning

Build a coherent emotional and functional association.

Attention requirement: longer, sustained, and voluntary. The study identifies Meaning as the task most reliant on the first 15–20 seconds of attentive viewing.

SMB-accessible formats: self-serve programmatic CTV, YouTube TrueView skippable, high-ceiling Meta video placements, Reddit and contextual podcast host reads.

Priority: attentive time supported by attentive reach. Invest in creative duration and opening-two-second craft.

D · Difference

Be seen as offering something others do not.

Attention requirement: longer and higher-quality than Salience, comparable to Meaning, and dependent on distinctive brand codes rather than rational argument.

SMB-accessible formats: non-skippable CTV, YouTube in-stream with masthead carry, premium Meta video in feed — in all cases with creative that foregrounds distinctive audio and visual assets.

Priority: attentive time, with budget reserved for repeat exposure of a consistent distinctive asset system.

Brand size modulates all three columns. Per the study, larger brands convert a given unit of attention into outcomes more efficiently than smaller brands; an SMB running below category penetration should budget for more attentive reach than the table above implies, and should expect compounded effects to emerge over a longer horizon. Source: dentsu, The Brand Reset, 2026 — Implications for Planners.

Self-assessment

The Signostic attention-quality checklist.

Ten yes-or-no questions derived from the five findings above. The checklist is designed to be completed in under five minutes by an in-house marketing lead. It does not replace an account-level audit, but it is a defensible first read on whether your media investment is positioned to capture the long-term effects the study describes.

Attention-quality self-diagnostic

Answer each item Yes or No. Tally the total yes responses at the end.

We measure YouTube, Meta, or TikTok video spend on at least one window longer than thirty days (for example, branded search query lift or 12-month recall).
At least one of our current creative units earns voluntary watch time past the five-second mark in reported analytics.
Our longest in-flight video unit is twenty seconds or shorter, or we have explicitly tested a shorter cut against the longer original.
We have run, or have a live quote for, at least one Connected TV placement via a self-serve DSP in the last twelve months.
Our hero creative is recognisable as our brand within the first two seconds, with sound off, on a muted mobile feed.
We track at least one long-term brand-health metric — unaided recall, branded search volume, consideration — in addition to short-window conversion metrics.
A cold viewer with no prior category knowledge could extract our value proposition from our hero creative in three seconds or less.
We use a consistent distinctive audio or visual signature across every paid video creative in flight this quarter.
Our reporting attributes at least some pipeline value to branded search and direct-site lift driven by upper-funnel video, rather than treating those channels as independent.
Our short-form video inventory is produced for platform-native requirements (vertical aspect ratio, first-frame brand mark, captions by default) rather than rebuilt from an existing broadcast cut.

7–10 Yes responses: your media operation is materially aligned with the attention-quality principles described in the study. Optimisation opportunities remain, but the structural pattern is sound.
4–6 Yes responses: a meaningful structural gap exists between your current reporting and the three-year value the study attributes to your spend. The findings above describe where to look first.
0–3 Yes responses: your account exhibits the textbook pattern the study documents. The five findings above are, in order, the places where the loop typically breaks.

References

Citations and source material.

The primary source for every statistic on this page is the dentsu report cited below. Items 2 through 8 are secondary references either used or commended within the dentsu study and are included here for completeness.

dentsu. The Brand Reset: Capturing Sales and Brand Equity from Attention. Global research study in partnership with Kantar and Lumen Research, with expert contribution from Dan White and advisory input from Les Binet. Published April 2026. Available at dentsu.com/reports/the_brand_reset_2026.
Kantar. Meaningful, Different, Salient (MDS) brand equity framework and the BrandZ sales simulator. Externally validated by the Marketing Accountability Standards Board.
Lumen Research. Attention measurement via eye-tracking methodology; long-term research partner of dentsu.
White, Dan. The Smart Marketing Book, The Smart Advertising Book, The Smart Branding Book. Co-developer of the BRANDZ brand-equity measurement system and methodological contributor to The Brand Reset.
Binet, Les & Field, Peter. Media in Focus: Marketing Effectiveness in the Digital Era. IPA, 2016. Foundational econometric case for long-term brand building, cited by the Brand Reset authors in support of Findings 01 and 02.
WARC. The Multiplier Effect: A CMO's Guide to Brand Building in the Performance Era. 2025. Source of the "doom loop" framing and multiplier model.
Jones, John Philip. When Ads Work: New Proof That Advertising Triggers Sales. Cited in the Brand Reset for the principle that there are no long-term effects without short-term effects.
IPA / Profit Ability 2, 2024; and The New Influencer Marketing Econometrics, IPA Effectiveness Conference, October 2025. Both cited by the Brand Reset authors as corroborating evidence for the long-term effectiveness of digital video.

Read the full report at dentsu.com ›