Audience Network Rewarded Video: How the Auction Works, What Creative Converts, and When to Stop

Most advertisers who set up Audience Network rewarded video do it wrong in the same specific way. They take a feed video ad — 15 seconds, branded hook, product demo, CTA — and push it into the rewarded placement without changing anything. Completion rates look fine. eCPM looks reasonable. And then post-view conversion rate sits 40% below what they see on Meta's own placements, and nobody can explain why.

The reason is direct: rewarded video is a completely different psychological context than a feed placement, and creative built for one performs poorly in the other.

TL;DR: Audience Network rewarded video is a voluntary, opt-in format inside third-party apps where users watch your video in exchange for an in-app reward. The opt-in screen is the most important frame. The auction runs on predicted eCPM adjusted for completion and conversion probability. It works well for app installs and subscription offers; it underperforms for direct-response ecommerce. Diagnose problems across three layers: completion rate, post-view conversion, and eCPM trend.

What Audience Network Rewarded Video Actually Is

Meta's Audience Network is the publisher side of Meta's advertising system — a network of third-party apps, primarily mobile games and utility apps, that display Meta-served ads. Publishers integrate the Audience Network SDK and choose which ad formats to offer: banner, interstitial, native, or rewarded video.

Rewarded video differs from every other format structurally: the user opts in. An in-app screen offers a trade — "Watch a 30-second video to earn 50 coins" — and the user chooses to accept. This is the opt-in mechanic, and it changes everything downstream.

Because the user chooses to watch, completion rates are dramatically higher than involuntary formats. Industry data from the Interactive Advertising Bureau's 2025 In-App Advertising Report shows rewarded video completion rates averaging 87-93% versus 51-64% for interstitial video. That completion signal feeds directly into Meta's delivery algorithm — the system treats rewarded inventory as higher quality because the engagement signals are reliable. That quality signal affects your auction position, your ad performance scores, and how quickly your campaign exits the learning phase.

For advertisers, the practical implication is that rewarded video gives you access to a user in an engaged, goal-directed mental state. They want something. They agreed to give you 15-30 seconds of attention to get it. That's not passive scrolling. The catch is that the same state that makes users highly engaged also makes them discerning. A viewer who opted in and then gets an irrelevant ad feels a stronger mismatch than a mid-scroll viewer — and that mismatch shows up in post-view metrics even when completion rate looks healthy.

See our guide on Meta Ads for App Install Campaigns for the broader placement strategy context.

How the Rewarded Video Auction and Delivery Work

The Audience Network auction for rewarded video runs on an effective CPM basis — not a raw bid basis. Your nominal bid is adjusted by Meta's predicted probability that a user will complete the video and take a downstream action, producing an eCPM that competes against other advertisers' adjusted eCPMs for the same impression.

Completion probability is weighted more heavily here than on feed placements. If your creative has a track record of high completion rates, your predicted eCPM is boosted before you even bid — meaning creative quality directly affects your cost per impression. Publisher floor prices are a second variable unique to Audience Network. Publishers set minimum eCPM floors for their rewarded inventory. A well-monetized mid-core gaming app with engaged 25-34 demographics sets higher floors than a utility app with mixed engagement. Your bid must clear the floor before entering the auction. If your campaign is underdelivering, the first diagnostic is whether your bids are clearing floors in your target app categories.

Meta's Audience Network developer documentation specifies that publishers can set both fixed and flexible floors that adjust based on demand pressure. In Q4 and peak flight periods, floors rise as more advertisers compete for premium rewarded inventory. Budget accordingly.

The learning phase on rewarded video is longer than on standard Meta placements. It requires approximately 50 optimization events in a 7-day window. For post-install event optimization (purchases, subscriptions), reaching 50 events takes longer and requires more budget before the algorithm stabilizes. Use our Ad Budget Planner to calculate the minimum spend needed to exit the learning phase within your campaign timeline.

For a deeper look at learning phase mechanics, see Mastering Meta Ads Learning Phase Optimization.

The Opt-In Screen: The Frame Most Guides Skip

The opt-in screen — the in-app prompt that appears before your video starts — determines whether a user watches your ad at all. And because it is controlled by the publisher, most advertisers assume they cannot influence it. That assumption is partially wrong.

Video lead-in. The first 1-2 seconds of your video are visible in some publisher implementations before the user confirms the opt-in. If that opening frame is visually interesting and contextually relevant to gaming contexts, opt-in rate increases. A face-forward UGC-style hook often outperforms a branded product shot here because it reads as human rather than commercial.

Reward alignment. When your offer has a natural reward structure — a free trial, a free level of your own app, a discount — mention it in the first 3 seconds of the video. Users who opted in for the publisher's reward and then hear about your offer's reward feel a consistent value-exchange narrative rather than a jarring transition.

Completion signal compounding. Meta's delivery system allocates more rewarded impressions to advertisers with better completion track records. Publishers who see higher opt-in and completion rates from a given advertiser's creative prefer that advertiser's inventory — it's a quality-signal flywheel. The advertisers who invest in rewarded-specific creative get progressively better placement.

For creative research on what rewarded-format structures are currently running in your category, AdLibrary's Ad Timeline Analysis shows which competitor creatives have been active longest — a proxy for which structures are working. Long-running creatives are rarely accidents.

See also AI UGC Video Ads Strategy for creative production approaches that translate well to mobile ad contexts.

When to Use Rewarded Video and When to Stop

Rewarded video earns its place in a media plan under specific conditions. Outside those conditions, it drains budget that looks acceptable in completion metrics while quietly underperforming on the metrics that matter.

Use rewarded video when your primary objective is app installs or subscriptions, your product has a natural reward structure (free trials, game-style mechanics, content tiers), you can produce dedicated rewarded creative, and your audience is broad enough for Audience Network delivery to find quality users. Custom audience retargeting lists under 100k constrain Audience Network delivery significantly — the format works best with lookalike audiences or broad interest targeting.

Stop rewarded video when completion rate is healthy but post-view CVR sits below 2% after 7+ days and 500+ completions — the creative is engaging but there's a fundamental product-context mismatch. Also stop when eCPM rises without corresponding volume improvement (you're winning lower-quality inventory as premium publishers fill their rewarded slots with higher-bidding competitors), or when your creative team cannot support dedicated rewarded production. Running repurposed feed assets in rewarded placements consistently underperforms the same budget on Meta's own placements.

For budget allocation modelling across placements, our Ad Budget Planner lets you model spend scenarios across different placement types. See also Automated Meta Ads Budget Allocation for rules-based budget management that responds to eCPM fluctuations automatically.

Campaign Setup: Four Levels, Four Decisions

Campaign level: Use App Installs, App Events, or Value Optimization. Rewarded video is not available for awareness, reach, or traffic objectives — it requires a downstream conversion event to optimize against.

Ad set — Placements: Select Audience Network manually and enable Rewarded Video specifically. Meta's Advantage+ Placements will include Audience Network automatically but won't differentiate between banner, interstitial, and rewarded. If you want isolated performance data for rewarded video, use manual placements and select only Audience Network Rewarded Video. Mixing formats in the same ad set makes it impossible to diagnose what's driving performance.

Ad set — Bidding: Start with Cost Cap rather than Lowest Cost. Rewarded inventory prices are volatile — floor prices spike during high-demand periods, and Lowest Cost bidding can overspend significantly during those windows. Set your Cost Cap at your target CPI plus 15% buffer and review every 3-4 days during the first two weeks.

Ad level — Creative specs: Meta Audience Network rewarded video requires 15-30 second MP4 or MOV files at 16:9 aspect ratio (some publishers support 9:16), minimum 720p resolution, with audio. The audio-on default is critical. Rewarded video is one of the few mobile placements where you can reliably assume audio will be heard — voice-over, music, and sound design all contribute in ways they cannot on feed where most users scroll muted.

For ad creative specification research on what formats competitors are running, see Best AI Tools for Ad Creative 2026 and the A Practical Guide to Competitor Ad Analysis.

eCPM Floors, Bidding Mechanics, and Real Cost Ranges

Understanding eCPM floors is not optional for rewarded video budgeting. Unlike Meta's own placements where your bid competes primarily against other advertisers, Audience Network adds a third variable: the publisher floor. Floor prices vary significantly by app category and user geography.

Based on Audience Network monetisation data from Meta's business help documentation: gaming apps (mid-core and hard-core) command the highest floors, often €8-€18 eCPM in Western European markets. Casual games sit in the €4-€10 range. Utility apps range from €2-€7. These are floors, not average costs — your actual winning bid is typically 10-40% above floor in competitive categories.

For budgeting purposes: if you're targeting a €2.50 CPI and your rewarded video completion rate is 88%, work backwards. At €2.50 CPI and a 3% post-completion install rate, you need an eCPM bid of approximately €7.50 to sustain that CPI at scale. If your target app category has floors above €7.50, either your CPI target needs adjustment or your conversion funnel needs improvement.

The Audience Saturation Estimator helps you model how audience size interacts with frequency and delivery costs — relevant here because rewarded video campaigns on narrow audiences hit frequency ceilings faster than broader placements, causing eCPM to rise as you compete for the same users repeatedly.

A useful external reference: Nielsen's 2025 Annual Marketing Report shows that mobile in-app advertising — including rewarded video — delivers 28% lower CPM-adjusted recall rates than social feed placements for brand objectives, but 34% higher post-view purchase intent for product categories with a clear value exchange. That data should inform your objective choice: rewarded video for conversion objectives, not awareness.

Measuring True Performance and Diagnosing Problems

Ad performance measurement on rewarded video requires separating three distinct conversion funnels: opt-in rate, completion rate, and post-completion conversion rate. Most advertisers look only at completion rate and post-view conversion rate — missing the opt-in layer because it isn't surfaced in standard Meta reporting.

Diagnosing completion rate below 75%: This is a creative problem. The user opted in — they want the reward. If they're dropping mid-video, the creative is losing them. Check the video drop-off report in Ads Manager. If drop-off concentrates in the first 5 seconds, the hook is failing. If drop-off concentrates at 10-15 seconds, the middle is losing momentum. Both are fixable with creative iteration — test the three variant structures (direct hook, problem hook, social proof hook) and run each to 500+ completions before drawing conclusions.

Diagnosing healthy completion with low CVR: When completion rate is strong (85%+) but post-completion install or purchase rate is below 2%, the creative strategy is delivering viewers who are not motivated buyers. This is a context mismatch — the rewarded video audience in this publisher category doesn't have the intent profile your offer requires. Test a different app category (switch from casual games to strategy games, or from games to utility apps) before overhauling the creative.

Diagnosing rising eCPM with flat conversion volume: When eCPM trend rises over 7-14 days without improvement in conversion volume, you're winning progressively lower-quality impressions. High-quality publishers fill their rewarded slots with the best-paying advertisers — as your campaign ages and you're no longer the newest, highest-quality creative in the auction, you drift toward lower-quality inventory. The fix: refresh your creative. A new creative resets the quality signal and can recover your position in the publisher quality stack.

For tracking which competitor creatives have been running longest — a sustained run being a strong signal of maintained eCPM position — use AdLibrary's Ad Timeline Analysis to monitor creative rotation in your competitive set. Frequent rotation (every 7-10 days) indicates a competitor burning through creatives due to creative fatigue. Infrequent rotation (30+ days) indicates a creative holding its quality position in the auction.

See Why Meta Ad Performance Is Inconsistent for the broader diagnostic framework, and Facebook Ads Workflow Efficiency for structuring your performance review cadence.

Frequency Management: Different Rules Than Feed

Frequency on Audience Network rewarded video requires a different management approach than on Meta's own placements. On Facebook and Instagram feed, high frequency primarily signals audience saturation. On Audience Network rewarded video, frequency accumulates differently because users actively request ad views to earn rewards — the same user may watch your ad multiple times voluntarily.

This creates a counterintuitive situation: high frequency on rewarded video does not automatically indicate creative fatigue or audience saturation. A user who watches your 30-second video three times may be genuinely interested in the offer. They chose to watch each time.

However, voluntary high frequency still creates a conversion problem. A user who has seen your install ad four times without installing is a non-converter. Continuing to show them your ad wastes budget. Recommended thresholds for rewarded video:

• Impressions per user per day: maximum 3
• Impressions per user per 7-day window: maximum 8
• After 8 impressions with no conversion: exclude from the rewarded placement and shift budget to feed retargeting with a differentiated offer

For lookalike audience expansion that reduces frequency pressure by adding fresh users, see Lookalike Audience Model 2026. For key performance indicator frameworks specific to app marketing, see Automated Meta Ads Budget Allocation.

Competitive Creative Research for Audience Network Placements

Most advertisers treat Audience Network as a black box — they set it up, run it, and never systematically research what competitors are doing on the placement. That gap is an advantage for anyone willing to close it.

The challenge: Audience Network creatives are served inside third-party apps, not in Meta's Ad Library (which indexes only Facebook and Instagram placements). The research approach is indirect. Track competitor ad accounts for video ad creatives matching rewarded video specifications — landscape 16:9, 15-30 seconds, audio-forward — and infer from ad duration and rotation frequency which ones are being used on Audience Network. Long-running 15-30 second landscape video ads that rotate infrequently are strong candidates for rewarded inventory.

What to study in those creatives: What's the hook structure? What's the reward narrative in the first 3 seconds? Is the CTA a direct install or a soft engagement invite? How does the creative frame the in-app reward context — does it acknowledge the gaming environment or ignore it?

AdLibrary's Ad Timeline Analysis tracks ad duration by creative across competitor accounts. Pair that with the methodology in Guide to Analyzing Competitor Ad Creative Strategies and A Practical Guide to Competitor Ad Analysis to build a structured view of which rewarded-format creative intelligence patterns are sustaining performance in your category.

For the creative testing cadence — building variant matrices from competitive research, testing methodically, refreshing before quality signal decay — see Facebook Ads Creative Testing Bottleneck for the structural framework.

Rewarded Video in the Broader App Marketing Stack

Audience Network rewarded video is most effective as one placement in a multi-channel app marketing stack. The users who convert on rewarded placements tend to be mid-funnel — they've had some exposure to your app or category, which is why the gaming-adjacent context resonates. Pure cold-audience campaigns on rewarded video with no prior touchpoints typically underperform against campaigns where the rewarded placement serves users who've seen your feed ads but haven't converted.

A practical sequencing framework:

1. Top of funnel: Feed video ads (Facebook and Instagram) for broad awareness, using lookalike audiences built from your highest-value existing users.
2. Mid-funnel: Audience Network rewarded video targeting users who've engaged with your Facebook/Instagram ads but not installed. The rewarded context gives them a low-friction second touchpoint.
3. Re-engagement: Feed retargeting with a differentiated offer (free trial, limited-time bonus) for users who've seen both feed and rewarded placements without converting.

This sequencing uses custom audience exclusions at each stage to prevent overlap and frequency waste. Configure your Audience Network ad sets to exclude existing users (via your app's SDK event list) and previous converters.

For the full app install campaign strategy across placement types, see Meta Ads for App Install Campaigns and Facebook Ads for Ecommerce Stores: The Stack That Scales for multi-placement campaign architecture principles.

For teams running systematic competitive creative research workflows at scale — pulling competitor ad timelines via API, classifying creative formats, tracking rotation frequency across accounts — AdLibrary's API access via the Business plan at €329/mo supports that pipeline with 1,000+ monthly credits. Manual media buyers running weekly competitive creative research on 5-10 accounts will find the Pro plan at €179/mo sufficient at 300 credits/month. Use the Creative Strategist Workflow as the operational template for that research cadence.

For dynamic creative testing specifically on rewarded placements — structuring creative variants by hook concept, offer framing, and CTA type — see Best AI Tools for Ad Creative 2026 and Shopify Competitor Revenue Analysis Guide for the competitive intelligence inputs that make variant briefs concrete rather than abstract.

Frequently Asked Questions

What is Audience Network rewarded video and how does it differ from standard video placements?

Audience Network rewarded video is a voluntary, opt-in video ad format served inside third-party mobile apps — typically games — where users choose to watch a video in exchange for an in-app reward. Unlike standard Audience Network interstitial or banner placements, the user initiates the ad experience. This opt-in mechanic produces completion rates of 85-95% versus 50-65% for interstitials. The auction runs on an eCPM basis adjusted for predicted completion and conversion probability, meaning creative quality directly affects your cost per impression, not just your cost per conversion.

How does the Audience Network rewarded video auction work?

The auction uses an eCPM-based system where your nominal bid is adjusted by Meta's predicted probability that a user will complete the video and take a downstream action. Higher predicted completion rates and post-view conversion rates raise your effective eCPM, so you can win inventory against higher nominal bids if your creative quality is superior. Publishers set floor prices through the Audience Network SDK, and your bid must clear that floor before entering the auction. Floor prices vary by app category and user geography — gaming apps in Western markets typically have higher floors than utility apps.

What video length and creative structure works best for rewarded video?

Rewarded video supports 15-30 second videos, with the opt-in screen being the most important frame. That screen must clearly communicate the reward value proposition because users who are not convinced skip before the video starts. Inside the video, the hook (first 3 seconds) should reinforce the reward context rather than selling immediately. Creative structured as problem → product solution → clear CTA outperforms hard-sell formats because users are in a reward-seeking mental state. Include a visible progress indicator to reduce mid-roll abandonment. Audio is on by default — design your creative assuming it will be heard.

When should you use Audience Network rewarded video and when should you avoid it?

Use rewarded video when your primary objective is app installs or subscriptions, your product has a natural reward structure, and you can produce dedicated rewarded-format creative. Avoid it when your offer has no alignment with a gaming or utility app context, when your creative team cannot build rewarded-specific assets, or when your audience targeting is too narrow for Audience Network delivery to find quality users. Stop the placement when completion rate is healthy but post-view CVR stays below 2% after 500+ completions — that signals a product-context mismatch that creative iteration cannot fix.

How do you diagnose poor performance on Audience Network rewarded video?

Diagnose across three layers. First, completion rate — below 75% signals a creative problem (hook failure or mid-video momentum loss). Second, post-view conversion rate — healthy completion with low CVR signals a product-context mismatch between your offer and the rewarded gaming audience. Third, eCPM trend — rising eCPM without volume improvement signals you are winning lower-quality inventory as higher-bidding competitors take premium publisher slots. Each layer requires a different fix: creative iteration, placement reconfiguration, or bid increase. Do not treat the three layers as interchangeable — applying the wrong fix to the wrong diagnostic layer wastes budget.

The Operational Bottom Line

Audience Network rewarded video is one of the most misused placements in mobile advertising — typically misused in the same direction. Advertisers treat it as a cheap impression source, push repurposed feed creative at it, see mediocre results, and conclude the placement doesn't work. The placement works. The creative and setup do not.

The teams that consistently outperform on rewarded video are the ones that treat the opt-in screen as the primary creative challenge, manage the eCPM floor mechanics actively rather than passively, and refresh creative before quality signal decay rather than after the metrics collapse. They run it as a deliberate mid-funnel placement, sequenced after feed exposure, targeted broadly enough for the algorithm to find quality users.

Before launching, spend 60-90 minutes on competitive creative research. Know which structures have been running 30+ days in your category. Know what reward narrative is working in the first 3 seconds. Know your competitors' rotation cadence so you can plan your production pipeline to match. That research changes your first creative brief from a blank-page exercise to a calibrated starting hypothesis.

For programmatic research at scale — pulling competitor ad timelines via API, classifying formats automatically, feeding outputs into briefing tools — the Business plan at €329/mo with API access gives your team the infrastructure to build that pipeline. For manual weekly research sweeps on your competitive set, the Pro plan at €179/mo at 300 credits/month covers the cadence. Either way, the Ad Timeline Analysis feature is the specific tool — track by duration, filter by format, sort by run length, and you have your research input in under an hour.