aCPM (Attention CPM): What It Is, How to Calculate It, and When to Use It

Attention CPM (aCPM) reframes media cost through a simple question: how much are you paying for exposure that people actually notice? Traditional CPM treats every served impression as equal. Viewability improved the hygiene bar, but it still can’t tell you if a human looked and processed the message. aCPM brings the human factor into the denominator by weighting cost against attention—often measured as attentive seconds.

Key takeaways

• aCPM is CPM adjusted by human attention, typically using attentive seconds per 1,000 impressions (APM) as the denominator.
• The common formula is aCPM ≈ CPM ÷ APM; some teams use cost per attentive second (CPAS) instead.
• aCPM is framework-dependent: different vendors define and model “attention” differently, so cross-vendor comparisons can mislead.
• Use aCPM alongside outcome metrics; attention is a quality signal, not a guaranteed sales driver.
• Benchmark aCPM by channel and format (CTV vs video vs display), and validate gains with brand lift, conversion lift, or MMM.

What is aCPM, in plain English?

Attention CPM (aCPM) is a media efficiency metric that expresses how much you pay per 1,000 impressions after adjusting for the amount of human attention those impressions actually deliver. Think of it as “CPM weighted by attention quality and duration.” If two placements have the same CPM but one generates more attentive viewing time, its aCPM will be lower—meaning you’re buying attention more efficiently.

What it is not:

• It’s not a replacement for CPM or outcome KPIs; it complements them.
• It’s not standardized; different vendors compute attention differently.
• It’s not a guarantee of performance—strong creative and fit-to-context still matter.

The formula and the units (read this first)

The most common expression connects CPM with APM (Attention per Mille):

• APM = attentive seconds per 1,000 impressions
• aCPM ≈ CPM ÷ APM

If APM is “2,000 attentive seconds per 1,000 impressions,” dividing CPM by 2.0 yields dollars per 1,000 attentive seconds. Some practitioners prefer to work directly with CPAS:

• CPAS (cost per attentive second) = Total Cost ÷ Total Attentive Seconds
• To convert CPAS to a “per 1,000 attentive seconds” expression, multiply CPAS by 1,000.

Vendors popularized attentive-time concepts and activation methods—Lumen, for example, frames attention volume and APM to enable “attention bargains,” where you bid toward attention rather than impressions alone, as described in their overview of attention technology and programmatic activation in 2023 (see Lumen’s explanations in Attention technology in advertising and Programmatic attention bidding).

Why units matter: aCPM has meaning only if you document the attention unit used (e.g., attentive seconds vs a probability score). Treat the denominator explicitly in briefs and reporting.

How “attention” is actually measured (and why it varies)

There’s no single industry standard, and methods differ by provider:

• Lumen Research: Models “attentive seconds per thousand impressions” by combining probability of gaze and dwell time, enabling attention-weighted buying (see Lumen’s attention technology overview, 2023).
• DoubleVerify (DV): Authentic Attention analyzes 50+ exposure and engagement signals—viewable time on screen, share of screen, audibility, and interactions—to produce an attention index used for benchmarking and optimization (see DV Global Insights 2024 and DV’s attention methodology explainer, 2024).
• Adelaide: Focuses on probability of attention as a media quality construct, producing an Attention Unit (AU) that it argues is more predictive of outcomes than duration alone (see Adelaide’s “Fundamentals of Attention Metrics,” 2024, and their critique of duration-only metrics, 2023).
• TVision (TV/CTV): Measures eyes-on-screen attention using opt-in panels and computer vision, reporting attentive seconds and related ratios for television and streaming (see TVision’s Why Attention page and State of Streaming 2025 report).

Because the definition of “attention” differs—time volume, probability quality, or composite indices—the aCPM denominator is vendor-specific. This is why you should avoid cross-vendor aCPM comparisons unless you normalize carefully.

Authoritative references:

• The idea that attentive time enables attention-weighted buying is discussed in Lumen’s 2023 write-ups on attention tech and activation (Lumen Research, 2023).
• DV’s multi-signal approach and outcome associations are documented in the DV Global Insights 2024 report (DoubleVerify, 2024).
• Adelaide’s AU rationale and outcome correlations are covered in their fundamentals and blog resources (Adelaide, 2023–2024).
• TVision publishes channel-specific attention benchmarks for CTV/TV (TVision, 2025).

Worked examples you can copy

Example A (display)

• Spend = $50,000
• Impressions = 10,000,000 → CPM = $5.00
• Attention partner reports APM = 2,000 attentive seconds per 1,000 impressions (i.e., 2.0 seconds per impression)
• aCPM = CPM ÷ APM = $5.00 ÷ 2.0 = $2.50 per 1,000 attentive seconds Interpretation: You’re effectively paying $2.50 for every 1,000 attentive seconds of exposure (as defined by your vendor).

Example B (video)

• Placement 1: CPM = $6.00; APM = 4,000 → aCPM = $6.00 ÷ 4.0 = $1.50 per 1,000 attentive seconds
• Placement 2: CPM = $4.00; APM = 1,000 → aCPM = $4.00 ÷ 1.0 = $4.00 per 1,000 attentive seconds Even with a higher sticker CPM, Placement 1 is more efficient on attention. This is the essence of finding “attention bargains.”

Alternative expression (CPAS): If Example A delivered 20,000,000 attentive seconds total, CPAS = $50,000 ÷ 20,000,000 = $0.0025 per second, equivalent to $2.50 per 1,000 attentive seconds.

CPM vs vCPM vs aCPM (and why this evolution matters)

• CPM: Cost per 1,000 served impressions. Blind to human attention.
• vCPM: Cost per 1,000 viewable impressions. Enforces a hygiene floor (pixels/time on screen) but still not proof of attention. The industry baseline for viewability—e.g., “≥50% of pixels in view for ≥1 continuous second” for display, and “≥50% for ≥2 continuous seconds” for video—comes from MRC/IAB guidelines that continue to anchor measurement standards (see IAB/MRC Retail Media Measurement Guidelines, 2024; Digital Video Served Impression Guidelines, 2018).
• aCPM: Cost per 1,000 “attention units” (commonly attentive seconds). Attempts to capture the portion of exposure likely to influence human cognition, with early evidence of closer linkage to outcomes in vendor analyses (see Adelaide’s fundamentals 2024 and DV Global Insights 2024).

At-a-glance comparison

For reference on viewability standards and extensions to newer formats, see the IAB/MRC retail media guidelines (2024) and the MRC’s video and AR/in-game measurement documents (2018–2024).

Using aCPM in practice: a step-by-step checklist

Before you compute aCPM

1. Choose your attention framework and vendor. Decide whether you’ll use attentive time (APM), a probability score (AU), or a composite index; document the unit.
2. Align on hygiene. Apply MRC/IAB viewability baselines so your attention numbers aren’t polluted by non-viewable impressions (see IAB/MRC Retail Media Measurement Guidelines, 2024).
3. Decide the efficiency expression. Will you report aCPM (per 1,000 attentive seconds) or CPAS (per-second)? Keep it consistent by channel.

During activation and optimization

1. Use pre-bid attention segments/optimizers where available and monitor their effect on both attention and outcomes. Providers document gains from attention-informed pre-bid segments and AI optimizers (see Adelaide’s pre-bid segments results, 2024; DV’s algorithmic attention optimizer announcement, 2023).
2. Run side-by-sides. Compare placements with similar CPMs but different APM/AU to identify attention bargains.
3. Establish floors. Set minimum attention thresholds (e.g., AU or APM) and redirect budget from persistently low-attention inventory.

Validation and governance

1. Tie attention to impact. Correlate attention shifts with brand lift or lower-funnel KPIs; where possible, run incrementality tests or feed attention variables into MMM. Industry trades report growing adoption but caution against declaring attention a buying standard prematurely (see buyer perspectives in 2024 coverage).
2. Re-benchmark by channel/format. Attention distributions differ in CTV vs social vs display; leverage channel-specific benchmarks (e.g., TVision’s State of Streaming 2025 for CTV).
3. Apply fraud and suitability controls. Attention can be distorted by made-for-advertising (MFA) sites and invalid traffic. Verification providers address MFA categorization and protection alongside attention (see DV’s MFA protection update, 2024).

Interpreting and benchmarking aCPM

• Channel specificity: Expect different APM ranges and aCPM baselines across formats. CTV and long-form video often deliver more attentive time than small display units; social in-feed has different patterns. Use channel-native benchmarks (e.g., TVision’s streaming reports for CTV, 2025).
• Vendor consistency: Keep planning, activation, and reporting within the same vendor model to avoid apples-to-oranges. If you must compare across vendors, create a translation layer (e.g., calibrate AU to attentive seconds using a holdout).
• Outcome linkage: Vendor studies show correlations between higher attention and improved brand/lower-funnel metrics, but there isn’t yet a universally accepted attention-to-outcome coefficient in open literature. Treat attention as a high-quality proxy and validate with brand lift studies, conversion lift, or MMM as you scale (see Adelaide’s case collections, 2024–2025; DV Global Insights 2024).

Risks, caveats, and what aCPM won’t solve

• Non-standardization: Different attention models (duration vs probability vs composite) mean the aCPM denominator varies. Avoid cross-vendor benchmarks unless normalized (see Adelaide’s critique of duration-only metrics, 2023).
• Outcome gap: Attention is necessary but not sufficient—poor creative or weak offers will still underperform.
• Fraud and MFA: High apparent “engagement” can be mechanical. Pair attention analytics with invalid traffic and quality filters (see DV’s MFA guidance, 2024).
• Privacy and signal loss: Favor approaches that rely on privacy-resilient exposure and interaction signals rather than deprecated identifiers (DV positions Authentic Attention accordingly in 2024 methodology materials).
• Channel heterogeneity: Do not transplant “good” aCPM thresholds across channels. Re-benchmark regularly.

Related and alternative metrics (and when to use them)

• APM (Attention per Mille): Attentive seconds per 1,000 impressions. Directly pairs with CPM to compute aCPM. Useful when you want an attention volume KPI; see Lumen’s explanations (2023).
• AU (Adelaide Attention Unit): Probability-of-attention media quality score. Use when you want a planning/optimization signal shown to correlate with outcomes in case studies; see Adelaide’s fundamentals (2024).
• CPAS (Cost per Attentive Second): Highly interpretable for finance partners. Convert to per-1,000 units for comparability if needed.
• vCPM (Viewable CPM): Maintain as a hygiene gate aligned to MRC/IAB; not a substitute for attention but still essential for quality control.
• qCPM or proprietary quality-weighted CPMs: Understand construction before using across partners; consider them directional.

Frequently asked questions

What is a “good” aCPM?

• It depends on channel, format, audience, and vendor model. Start by benchmarking within your campaigns and portfolios, then set directional targets. Use lift studies or MMM to connect aCPM improvements to actual outcomes before hard-coding goals.

Can I use aCPM to compare different vendors or platforms?

• Not directly. Because attention definitions vary, aCPM is framework-dependent. Keep comparisons within a vendor’s ecosystem or create a normalization study to translate units.

Does creative quality affect aCPM?

• Yes. Better creative typically increases attentive time or the probability of attention, reducing aCPM at the same CPM. Treat creative testing as part of your attention optimization.

How should I report aCPM to finance stakeholders?

• Consider CPAS (cost per attentive second) alongside aCPM. The per-second unit can be easier to grasp and ties neatly to budget math. Always specify the attention unit and source.

Should I buy on attention directly?

• Some buyers use pre-bid attention segments or attention-indexed PMPs and report efficiency gains, but leading brands still caution that it’s too early to make attention a universal buying currency. Pilot, measure, and scale where it proves out in your business (see trade coverage of early adoption in 2023 and buyer perspectives in 2024).

Sources and further reading

• Lumen Research explains how attentive seconds and APM enable attention-weighted buying in their 2023 articles: “Attention technology in advertising” and “Programmatic attention bidding.”
• DoubleVerify details its multi-signal Authentic Attention approach and outcome associations in the “DV Global Insights 2024” report and related explainers.
• Adelaide outlines the probability-of-attention AU metric and its outcome links in “The fundamentals of attention metrics” (2024) and critiques duration-only methods in 2023 blog posts.
• TVision provides eyes-on-screen TV/CTV attention measurement and streaming benchmarks in “Why Attention” (updated 2024/2025) and the “State of Streaming 2025” report.
• For viewability baselines and evolving standards, see IAB/MRC Retail Media Measurement Guidelines (2024), MRC Digital Video Served Impression Measurement Guidelines (2018), and format extensions like AR and in-game (2022–2024).

Cited materials (anchor references within the article):

• The industry’s pixel/time viewability definitions are documented in the IAB/MRC Retail Media Measurement Guidelines (2024) and the MRC’s Digital Video Served Impression Measurement Guidelines (2018).
• Lumen’s framing of attention volume and attention-weighted activation appears in their 2023 posts on attention technology and programmatic attention bidding.
• DV’s attention index and outcome associations are reported in the DV Global Insights 2024 report, and DV’s blog explains how attention metrics demonstrate premium inventory value (2024).
• Adelaide’s AU methodology and its stance on probability over duration are in the 2024 fundamentals and the 2023 “debunking duration-based metrics” piece; they also share pre-bid segment results (2024) and an outcomes case study collection (2024–2025).
• TVision’s State of Streaming 2025 provides CTV attention benchmarks, reinforcing the need for channel-specific baselines.

---

References (selected, linked above):

• IAB/MRC Retail Media Measurement Guidelines (2024)
• MRC Digital Video Served Impression Measurement Guidelines (2018)
• Lumen Research (2023): attention technology; programmatic attention bidding
• DoubleVerify (2024): Global Insights; attention methodology and MFA protection update (2024)
• Adelaide (2023–2025): fundamentals; debunking duration-only; pre-bid results; outcomes guide
• TVision (2025): State of Streaming; Why Attention