In-Game Advertising Measurement: From Exposure to Outcomes

In-game advertising measurement fails when every format is forced into one impression count. A static surface in a 3D world, an opt-in video, an audio message, a portal and a playable brand integration create different player opportunities and require different evidence.

The measurement architecture should follow a simple rule: do not let evidence from one stage impersonate the next. Delivery is not viewable exposure. Exposure is not attention. Attention is not interaction. Interaction is not brand lift. Attributed conversion is not necessarily incremental return.

Executive Summary

1. 1Classify the format first.Display, video, audio, portals and custom integrations have different delivery and player-action mechanics.
2. 2Preserve the evidence ladder.Eligibility, delivery, exposure, attention, interaction and outcomes need separate definitions and denominators.
3. 3Validate before valuing.Geometry, duration, technical quality, invalid traffic and data completeness precede performance interpretation.
4. 4Add a counterfactual for causal claims.Attribution can connect events; incrementality explains what happened because the advertising existed.

Start with the advertising format#

The IAB Gaming Measurement Framework organizes gaming advertising across display, video, audio and custom formats, with baseline and additional metrics. That is more useful than treating gaming as one placement type.

Format	Primary delivery unit	Player-context question	Additional evidence
Intrinsic display	Served and viewable impression	Was the surface plausibly visible under the disclosed rule?	Exposure duration, attention, recall or action
In-game video	Start, viewable impression or qualified view	Was viewing autoplay, opt-in, rewarded or interrupted?	Watch duration, audibility, completion and outcome
In-game audio	Eligible audible exposure	Was audio active, audible and presented in the relevant context?	Listening evidence, recall and outcome
Portal or destination	Eligible portal exposure and successful arrival	Did the player intentionally move into the advertiser destination?	Destination quality, qualified play and return
Custom integration	Defined branded-context exposure or participation	What part of gameplay, narrative or utility carried the brand?	Attention, meaningful action, lift and business outcome

The gaming brand activations hub covers the wider category. This page owns the paid in-game measurement architecture; the brand activation measurement playbook owns the complete campaign journey across product, media and outcomes.

The seven-layer measurement architecture#

1. 1Eligible
   • ·Player policy
   • ·Market
   • ·Device
   • ·Targeting
2. →2Delivered
   • ·Creative served
   • ·Unit valid
   • ·Technical success
   • ·IVT filtered
3. →3Exposed
   • ·Visibility
   • ·Duration
   • ·Angle
   • ·Audibility
4. →4Attended
   • ·Data signals
   • ·Visual or audio
   • ·Panel
   • ·Validation
5. →5Acted
   • ·Click
   • ·View
   • ·Teleport
   • ·Gameplay action
6. →6Changed
   • ·Recall
   • ·Consideration
   • ·Search
   • ·Conversion

Financial contribution and ROI sit after this diagram because they need an incrementality model and complete cost base, not another delivery metric.

Eligibility is part of the denominator#

Not every player in a game can or should receive every advertisement. Audience selection, market, device, age eligibility, platform policy, frequency and inventory availability can all reduce the eligible population.

Report at least four populations separately:

1. players active in the relevant game context;
2. players eligible to receive the advertising;
3. eligible players to whom the creative was successfully delivered;
4. delivered players who met the viewable-exposure rule.

If an ad unit shows fallback content to an ineligible player, that player was in the environment but did not receive the advertiser’s creative. Using total game visits as ad reach would inflate the denominator and the result.

Exposure rules must be format-specific#

Roblox’s immersive-ad documentation demonstrates the specificity required. Its current image and autoplay-video impression rules include duration, viewport share, pixel visibility and viewing angle. Click-to-play video and portal ads use different qualifying actions: video viewing and successful teleport.

The strategic lesson is not to copy one platform’s thresholds into another environment. It is to demand a disclosed rule that matches the format.

Evidence field	Why it matters	Failure mode
Creative served	Confirms the advertiser asset was delivered	Counting fallback or empty inventory
Geometry and occlusion	Confirms the surface had plausible visibility	A surface exists but faces away or is blocked
Duration	Separates momentary intersection from qualified exposure	A frame becomes a full impression
Player and camera state	Tests whether the user was active in the relevant context	Loading, idling or menu time is treated as exposure
Audibility	Establishes an opportunity to hear audio or video	Muted delivery is interpreted as heard
Invalid-traffic control	Removes or labels automated and manipulated activity	Synthetic delivery inflates performance

The intrinsic in-game guidelines focus on impression, viewable impression and audience measurement. They do not make delivery a synonym for interaction or outcome. That boundary is a feature, not a limitation.

Viewability is not attention#

Viewability establishes a technical opportunity to notice. Attention attempts to estimate whether notice occurred. The IAB, MRC and CIMM framework identifies several attention-method families, including data signals, visual or audio tracking, physiological observation, and panel or survey approaches.

Choose the method from the decision and risk. A directional creative test may use validated behavioral signals. A high-value claim about cross-format attention may require a stronger and independently validated design.

ROLearn’s measuring attention, not traffic framework shows how to qualify behavioral proxies without labeling every active minute as conscious branded attention.

Interaction needs an eligible denominator#

A click, opt-in view, teleport, quest completion or item use can indicate deliberate behavior. The rate becomes meaningful only when the eligible denominator is stated.

Action	Useful denominator	Interpretation boundary
Click or expand	Players with a valid exposed interactive unit	Interest in the unit, not necessarily the brand
Qualified video view	Players who received the relevant video opportunity	Viewing under the format rule, not persuasion
Portal arrival	Players exposed to an eligible portal	Intentional destination entry, not a qualified destination experience
Branded mechanic completion	Eligible starters of the mechanic	Participation depth, not brand lift
Item use later	Players who acquired the item and had an eligible later context	Persistence or utility, not incremental purchase

Always show the funnel beneath the headline. A high teleport count can coexist with weak destination load success or immediate exits. Optimizing the upstream action alone can move cost into a broken downstream experience.

Outcome measurement begins with exposure quality#

The MRC Outcomes and Data Quality Standards make an important governing point: outcome measurement depends on accurate delivery and exposure evidence. A conversion model cannot repair an exposure log that assigned treatment to the wrong people.

For brand outcomes, define the population, exposure, instrument, comparison, sample, field dates and uncertainty. For behavioral outcomes, define the event, identity or matching rule, attribution window and exclusions. For incremental claims, add a randomized holdout, credible matched comparison, staggered design or another defended counterfactual.

Outcome claim	Minimum evidence	Safe conclusion
Attributed site visit	Eligible exposure plus stated identity and time window	The visit matched the attribution rule
Brand lift	Defined exposure and credible exposed-versus-control outcome design	The measured outcome differed under the study design
Incremental conversion	Observed outcome plus defended counterfactual	Estimated conversions caused by the advertising
ROI	Incremental financial contribution and complete campaign cost	Return under the stated financial model

Keep earned media value in a separate ledger#

An in-game campaign can create creator videos, player clips, social posts and press coverage. That earned distribution matters, but it is not paid ad delivery and should not be merged into impression reporting.

ROLearn’s OMNI-EMV earned media value framework values attributable attention in each channel’s native unit, applies evidence and overlap controls, and reports a bounded media-equivalence estimate. Brand lift, conversion, revenue and ROI remain separate.

The buyer’s measurement specification#

Before media or production is contracted, the brief should require:

• format and player interaction model;
• eligible population and policy exclusions;
• delivery, viewability and exposure definitions;
• attention method and validation, if attention will be claimed;
• interaction events and denominators;
• invalid-traffic, data-quality and reconciliation policy;
• attribution windows and identity rules;
• outcome design and counterfactual;
• market, platform, device and creative breakdowns;
• raw-versus-modeled fields, uncertainty and correction policy;
• paid, owned and earned distribution separation;
• total campaign cost boundary for any ROI claim.

How this report was built

Format

Display, video, audio, portal and custom integrations classified before metrics are selected.

Exposure

Eligibility, delivery, geometry, duration, audibility, player state and invalid traffic documented.

Inference

Viewability, attention, interaction, attribution and incrementality labeled as different evidence levels.

Reporting

Paid delivery, earned media value, brand outcomes, business outcomes, uncertainty and cost shown in separate ledgers.

Read the full methodology →

Framework reviewed 12 August 2026. Platform definitions, eligibility policy and standard versions must be frozen for every campaign report.

Frequently asked questions#

What is in-game advertising measurement?#

In-game advertising measurement is the process of defining and validating delivery, viewable exposure, attention, interaction and outcomes for advertising delivered inside or through game environments.

How is an in-game ad impression measured?#

The rule depends on the format and platform. A defensible impression requires eligible delivery plus disclosed visibility, duration, geometry and invalid-traffic conditions; it should not be inferred from a player merely entering the game.

Is in-game ad viewability the same as attention?#

No. Viewability establishes a technical opportunity to see or hear. Attention requires additional validated evidence that the ad was likely noticed, and attention still does not prove a brand or business outcome.

How do you measure a custom brand integration in a game?#

Define the branded context, eligible player population, exposure rule and meaningful actions before launch, instrument the experience, measure outcomes with an appropriate comparison, and report off-platform earned distribution separately. The broader brand impact in games framework shows how exposure connects to memory and decision outcomes.

How should brands calculate ROI from in-game advertising?#

Use incremental financial contribution under a credible counterfactual and divide by the complete campaign cost. Do not use impressions, attention, attributed revenue or earned media value as a substitute for incremental return.