Forward to: Analytics Team

Measurement &
Attribution Workflows

Ten agent workflows for the platform's analytics team — multi-touch attribution enrichment, incrementality measurement, media mix modeling inputs, cross-channel conversion intelligence, advertiser benchmarking, campaign lift measurement, audience quality scoring for attribution, viewability-to-conversion analysis, geo-attribution enhancement, and predictive ROAS modeling across 102M domains.

1Multi-Touch Attribution Domain Enrichment

AI agent enriches the platform's multi-touch attribution model by adding domain-level context to every touchpoint — classifying each publisher impression by IAB Category, PageRank quality tier, and Persona type to understand which types of publisher interactions drive the most conversions.

1
Enrich Attribution Touchpoints with Domain Signals
/about /blog IAB Categories OpenPageRank Personas
ATTRIBUTION ENRICHMENT — TOUCHPOINT ANALYSIS ══════════════════════════════════════════════ CONVERSIONS ANALYZED: 34.5M attributed conversions this quarter TOUCHPOINTS ENRICHED: 234M impressions with domain-level signals AVG TOUCHPOINTS PER CONVERSION: 6.8 enriched touchpoints ATTRIBUTION VALUE BY PUBLISHER CONTEXT: Review Sites (IAB: Shopping > Reviews): Avg position in path: 3rd (mid-funnel) Attribution credit: 28% of conversion value Key insight: Review site touchpoints increase CVR by 67% News Sites (IAB: News, PageRank >7): Avg position in path: 1st (upper-funnel awareness) Attribution credit: 12% of conversion value Key insight: Premium news builds awareness for later conversion Social/Entertainment (IAB: Entertainment): Avg position in path: 2nd (consideration) Attribution credit: 18% of conversion value Key insight: Entertainment context drives aspirational interest Retailer Sites (IAB: Shopping, high intent): Avg position in path: Last (conversion) Attribution credit: 42% of conversion value Key insight: Final touchpoint on retailer domains closes the sale
2
Optimize Budget Allocation Based on Attribution
Agent translates enriched attribution insights into budget allocation recommendations — shifting spend toward publisher contexts that contribute most to conversion paths while maintaining upper-funnel awareness on premium publishers.
Attribution Insight
Review Site Impact — Review site touchpoints (wirecutter.com, rtings.com, pcmag.com) appear in 67% of high-value conversion paths. Attribution credit: 28% of total. These mid-funnel touchpoints are undervalued in last-click attribution but critical for conversion. Recommendation: Increase review site CPM bids by 35%.
INCREASE BID +35% — Review sites undervalued in attribution model
Attribution Insight
Premium News Awareness — High-PageRank news sites (nytimes.com, bbc.com) appear as first touchpoint in 45% of conversion paths. Low direct attribution credit (12%) but removing them decreases overall conversion volume by 23%. Recommendation: Maintain awareness spend but optimize frequency.
MAINTAIN SPEND — Awareness touchpoints critical for conversion funnel
3
Output: Attribution-Optimized Budget Allocation

MTA ENRICHMENT — BUDGET OPTIMIZATION

ATTRIBUTION-OPTIMIZED ALLOCATION: ══════════════════════════════════════════════ Review sites: +35% budget (undervalued mid-funnel impact) Premium news: Maintain (awareness critical for conversion paths) Entertainment: -10% budget (high CPA, low attribution credit) Retailer sites: Maintain (highest last-touch conversion) Revenue impact: +$12.3M/quarter from attribution-optimized spend ROAS improvement: +28% Enrichment: IAB Categories, PageRank, Personas, domain path analysis

2Incrementality Measurement Engine

AI agent measures true incrementality of ad platform campaigns by analyzing domain intelligence to create control groups — matching exposed and unexposed users by their domain visit profiles (IAB Categories, Personas) rather than random assignment, improving measurement accuracy.

1
Build Domain-Matched Control Groups
/products IAB Categories Personas Countries
INCREMENTALITY — DOMAIN-MATCHED CONTROL GROUPS ══════════════════════════════════════════════ CONTROL GROUP METHODOLOGY: Traditional (random): Control accuracy 72% (unmatched profiles) Domain-matched: Control accuracy 94% (matched by domain profile) Matching criteria: IAB Category distribution: Within 5% match Persona mix: Within 8% match Geographic distribution (Countries): Within 3% match PageRank tier distribution: Within 5% match INCREMENTALITY RESULTS (Nike Q1 2026 campaign): Exposed group CVR: 4.2% (234K conversions) Control group CVR: 2.8% (156K conversions) True incremental lift: +50% (78K incremental conversions) Incremental ROAS: 5.4x Traditional method would have shown: +72% lift (inflated by 44%) Domain-matched method: +50% true lift (accurate measurement)
2
Validate Control Group Quality with Domain Matching
Measurement Signal
Nike Q1 2026 — Domain-Matched Control — Control group domain profile matches exposed group within 5% on all dimensions: IAB Sports 34% (vs 35%), Fashion 23% (vs 22%), Health 18% (vs 19%). Persona match: Active Lifestyle 45% (vs 44%). Geographic match: US 67% (vs 68%). Match quality: 94% (vs 72% for random assignment). True incremental lift: +50% (not inflated +72% from random).
HIGH-QUALITY MATCH — 94% control accuracy, true +50% incremental lift
Measurement Signal
Traditional Random Control — Inflated Results — Random control group had mismatched domain profiles: Tech-heavy control (28% IAB Tech) vs sports-heavy exposed (34% IAB Sports). This mismatch inflated measured lift by 44% (showing +72% instead of true +50%). Domain-matched methodology corrects this bias and gives advertisers accurate incrementality they can trust for budget decisions.
BIAS DETECTED — Random control inflated by 44%, domain matching corrects
3
Output: Incrementality Dashboard

INCREMENTALITY — MEASUREMENT REPORT

INCREMENTALITY MEASUREMENT RESULTS: ══════════════════════════════════════════════ Campaigns measured: 4,567 this quarter Avg true incremental lift: +38% (vs inflated +62% from random controls) Control group accuracy: 94% (domain-matched) Advertiser confidence: +56% in platform measurement reports Revenue from measurement-driven retention: $23.4M/year

3Media Mix Modeling Domain Inputs

AI agent generates media mix modeling inputs by classifying ad platform campaign impressions by publisher domain characteristics — providing advertisers with channel-level performance data segmented by content type, quality tier, and audience composition for MMM analysis.

1
Generate MMM-Ready Domain-Level Data
/about IAB Categories OpenPageRank Personas Countries
MEDIA MIX MODEL INPUTS — DOMAIN SEGMENTATION ══════════════════════════════════════════════ PLATFORM IMPRESSION BREAKDOWN FOR MMM: Premium Publishers (PageRank >7): 28% of impressions | ROAS: 6.8x Mid-Tier Publishers (PageRank 4-7): 45% of impressions | ROAS: 4.2x Long Tail (PageRank <4): 27% of impressions | ROAS: 2.1x By IAB Category: News: 23% | ROAS: 4.8x (brand awareness contribution) Shopping: 34% | ROAS: 7.2x (high intent) Entertainment: 18% | ROAS: 3.1x Tech: 12% | ROAS: 5.4x Other: 13% | ROAS: 2.8x MMM EXPORT FORMAT: Weekly aggregates by PageRank tier, IAB, Country
2
Segment Platform Data for MMM Granularity
MMM Signal
Premium Publisher Segment — ROAS 6.8x — Domain-enriched MMM input: 28% of platform impressions on PageRank >7 publishers delivering 6.8x ROAS. IAB breakdown: News (45%), Business (22%), Technology (18%). This segment shows the platform's strongest performance and should be highlighted in MMM models as "premium programmatic" — a distinct channel from "long tail programmatic" (2.1x ROAS).
PREMIUM SEGMENT — 6.8x ROAS, separate from long tail in MMM model
MMM Signal
Shopping Context Segment — Highest Intent — 34% of platform impressions on IAB: Shopping domains delivering 7.2x ROAS. This segment captures high-intent users on retailer and review sites. When separated in MMM models, the platform's shopping-context performance exceeds Google Shopping (6.1x) and Meta Product Ads (4.8x). Key competitive advantage for MMM presentations.
SHOPPING CONTEXT — 7.2x ROAS, outperforms Google/Meta in MMM
3
Output: MMM Data Quality Report

MMM INPUTS — DATA QUALITY REPORT

MMM INPUT DATA QUALITY: ══════════════════════════════════════════════ Impressions classified: 98.4% (with domain-level signals) Granularity: Weekly by PageRank tier, IAB category, Country Historical depth: 8 quarters of enriched data Advertiser MMM satisfaction: +45% vs non-enriched data MMM-attributable platform value: +23% higher than pre-enrichment

4Cross-Channel Conversion Intelligence

AI agent tracks cross-channel conversion paths by monitoring advertiser domain activity — /products page visits, /login activity, and /pricing page engagement — to understand how ad platform campaigns interact with other marketing channels in driving purchases.

1
Map Cross-Channel Conversion Journeys
/products /login /pricing IAB Categories Personas
CROSS-CHANNEL CONVERSION PATHS ══════════════════════════════════════════════ MOST COMMON CONVERSION PATHS (with platform touchpoint): Path 1 (23%): Google Search → Platform Retarget → Purchase Platform contribution: Last touch, conversion driver Path 2 (18%): Social (Meta) → Platform Retarget → Review Site → Purchase Platform contribution: Mid-funnel, kept user in consideration Path 3 (15%): Platform Prospecting → Google Search → Purchase Platform contribution: First touch, drove awareness Path 4 (12%): Email → Platform Commerce Ad → Purchase Platform contribution: Reinforcement, aided conversion CROSS-CHANNEL SYNERGY: Platform + Google: +34% combined ROAS vs either alone Platform + Social: +28% combined ROAS vs either alone Platform + Email: +45% combined ROAS vs either alone
2
Quantify Platform Synergy with Other Channels
Synergy Signal
Platform + Google — +34% Combined ROAS — Domain intelligence shows users exposed to platform commerce ads AND Google search ads convert at 34% higher ROAS than either channel alone. Path analysis: Platform impression on review site (mid-funnel) followed by Google search (high-intent) is the highest-converting 2-touch sequence. Platform plays critical mid-funnel role that Google alone cannot fill.
CHANNEL SYNERGY — Platform + Google = +34% ROAS vs either alone
Synergy Signal
Platform Assisted Conversions — 25% Under-Counted — Domain path analysis reveals 25% of platform-influenced conversions are attributed to other channels in last-click models. Platform touchpoints appear in mid-funnel (review sites, comparison sites) but Google Search or direct gets last-click credit. Multi-touch attribution shows platform's true contribution is 67% of conversions, not 42%.
UNDER-COUNTED — 25% of platform conversions attributed to other channels
3
Output: Cross-Channel Value Report

CROSS-CHANNEL — CONVERSION VALUE

PLATFORM CROSS-CHANNEL CONTRIBUTION: ══════════════════════════════════════════════ Conversions with platform in path: 34.5M (67% of total) Avg platform position in path: 2.3 (mid-funnel critical role) Cross-channel ROAS synergy: +36% avg lift when platform present Standalone attribution: 42% of conversions Assisted attribution: 25% additional (under-counted without MTA) Signals: /products, /login, /pricing domain visits, IAB, Personas

5Advertiser Benchmarking Intelligence

AI agent generates advertiser performance benchmarks by analyzing domain intelligence across IAB Categories — providing advertisers with context on how their ad platform campaigns perform relative to their vertical, allowing for data-driven optimization recommendations.

1
Build Category Benchmarks from Domain Data
/products IAB Categories OpenPageRank Countries
ADVERTISER BENCHMARKING — BY IAB CATEGORY ══════════════════════════════════════════════ CATEGORY BENCHMARKS (Q1 2026): Category Avg CTR Avg CPA Avg ROAS Top Quartile Electronics 2.1% $12.40 6.8x 9.2x Fashion 1.8% $8.90 5.4x 7.8x Home & Garden 1.4% $15.20 4.2x 6.1x Beauty 2.4% $6.80 7.1x 10.2x Sporting Goods 1.6% $11.50 4.8x 6.9x Pet Supplies 1.9% $9.20 5.6x 8.1x Grocery 2.8% $4.50 8.4x 12.1x METHODOLOGY: Based on 47,234 advertisers segmented by IAB Category
2
Generate Advertiser-Specific Optimization Recommendations
Benchmark Signal
Allbirds — Below Benchmark (Fashion) — Current ROAS: 3.8x vs category benchmark 5.4x (top quartile: 7.8x). Domain intelligence diagnosis: Allbirds campaigns over-index on entertainment sites (34% of spend) where fashion CTR is 1.1% vs review sites where fashion CTR is 3.4%. Recommendation: Shift 20% of budget from entertainment to review/shopping contexts to reach benchmark.
BELOW BENCHMARK — 3.8x vs 5.4x avg, shift spend to review contexts
Benchmark Signal
Sephora — Top Quartile (Beauty) — Current ROAS: 10.2x vs category benchmark 7.1x (top quartile: 10.2x). Sephora is the benchmark leader in Beauty. Domain intelligence shows Sephora's advantage: 78% of spend on beauty/lifestyle review sites with optimal persona alignment. Share this playbook anonymously with below-benchmark beauty advertisers to lift category performance.
TOP QUARTILE — 10.2x ROAS, category leader, playbook for others
3
Output: Benchmarking Report

ADVERTISER BENCHMARKS — QUARTERLY REPORT

BENCHMARKING IMPACT: ══════════════════════════════════════════════ Advertisers benchmarked: 47,234 across 400+ IAB categories Below-benchmark advertisers identified: 12,345 (optimization opportunity) Optimization recommendations generated: 8,901 ROAS lift from benchmark-driven optimization: +23% Advertiser retention improvement: +18% (data-driven confidence)

6Campaign Lift Measurement

AI agent measures campaign lift by comparing advertiser domain activity before and after ad campaign exposure — tracking /products page visits, /pricing page engagement, and /login activity as proxy metrics for brand lift and purchase intent.

1
Measure Domain Activity Lift Post-Exposure
/products /pricing /login OpenPageRank
CAMPAIGN LIFT — DOMAIN ACTIVITY MEASUREMENT ══════════════════════════════════════════════ NIKE Q1 2026 CAMPAIGN LIFT RESULTS: /products page visits (exposed vs control): Exposed: +34% more visits to nike.com/products Control: baseline (no change) Lift: +34% product page engagement /pricing page visits: Exposed: +28% more visits to nike.com pricing pages Control: baseline Lift: +28% purchase intent /login activity: Exposed: +18% more account logins Control: baseline Lift: +18% account engagement COMPOSITE BRAND LIFT SCORE: +26.7% (weighted average)
2
Correlate Lift with Publisher Domain Quality
Lift Signal
Nike /products Page Lift — +34% on Premium Publishers — Exposure on PageRank >7 publishers drove +34% lift in nike.com /products page visits. Premium publisher context creates stronger brand lift than mid-tier (mid-tier: +18% lift). /pricing page lift was +28% (indicating purchase intent). Campaign on premium publishers generates 1.9x the brand lift per impression vs long tail.
PREMIUM LIFT — +34% product page visits from premium publisher exposure
Lift Signal
Nike /login Lift — +18% Account Engagement — Campaign exposure drove +18% increase in nike.com /login activity (account sign-ins). This is a deeper engagement metric than product page visits — indicating the campaign drove not just browsing but account-level commitment. /login lift is the strongest predictor of eventual purchase (0.78 correlation with conversion).
DEEP ENGAGEMENT — +18% login lift, strongest purchase predictor
3
Output: Campaign Lift Report

CAMPAIGN LIFT — MEASUREMENT REPORT

CAMPAIGN LIFT MEASUREMENT: ══════════════════════════════════════════════ Campaigns measured for lift: 2,345 this quarter Avg product page lift: +29% Avg pricing page lift: +22% Avg login activity lift: +15% Composite brand lift: +22% average across campaigns Advertiser confidence in the platform: +34% improvement

7Audience Quality Scoring for Attribution

AI agent scores audience quality within attribution paths by analyzing the domain profiles of converting users — determining which audience segments (by Personas and IAB affinity) generate the highest-value conversions for more accurate budget allocation.

1
Score Audience Quality by Conversion Value
/products Personas IAB Categories OpenPageRank
AUDIENCE QUALITY — CONVERSION VALUE ANALYSIS ══════════════════════════════════════════════ AUDIENCE QUALITY TIERS (by conversion value): Premium Buyers (Persona: Luxury, PageRank >7 sites): Avg order value: $289 | CVR: 3.8% | LTV: $1,234 Active Shoppers (Persona: Active Buyer, Review sites): Avg order value: $134 | CVR: 4.2% | LTV: $567 Casual Browsers (Persona: Browser, Entertainment sites): Avg order value: $67 | CVR: 1.2% | LTV: $189 Deal Seekers (Persona: Value Seeker, Coupon sites): Avg order value: $45 | CVR: 5.6% | LTV: $123 INSIGHT: Premium Buyers are 10x more valuable by LTV than Deal Seekers
2
Compare Audience Tiers by Lifetime Value
Quality Signal
Premium Buyers — 10x LTV vs Deal Seekers — Domain intelligence confirms massive quality gap: Users visiting luxury/high-PageRank domains (Persona: High-Net-Worth) have LTV of $1,234 vs $123 for deal/coupon site visitors (Persona: Value Seeker). CPAs are similar ($12 vs $8) but LTV makes Premium Buyers 10x more valuable. Attribution should weight Premium Buyer conversions 10x in budget allocation.
10x LTV GAP — Premium Buyers $1,234 LTV vs Deal Seekers $123
Quality Signal
Deal Seekers — High CVR Misleads Standard Attribution — Deal Seekers show 5.6% CVR (highest of all tiers) but $45 AOV and $123 LTV. Standard attribution overvalues this audience because it optimizes for conversion count, not value. Quality-adjusted attribution reveals Deal Seekers generate only 14% of revenue despite 28% of conversions. Reduce bid multiplier by 30%.
MISLEADING CVR — High conversions but low value, reduce bids 30%
3
Output: Quality-Adjusted Attribution

AUDIENCE QUALITY — ATTRIBUTION ADJUSTMENT

QUALITY-ADJUSTED ATTRIBUTION: ══════════════════════════════════════════════ Revenue from Premium Buyer audiences: $89M/quarter (38% of total) Revenue from Active Shoppers: $67M/quarter (29% of total) Revenue from Casual Browsers: $45M/quarter (19% of total) Revenue from Deal Seekers: $34M/quarter (14% of total) Quality-adjusted ROAS: +23% higher when weighting by LTV Budget reallocation: Shift 15% from Deal Seekers to Premium Buyers

8Viewability-to-Conversion Analysis

AI agent correlates publisher domain viewability metrics with conversion rates — analyzing how PageRank, IAB Category, and domain quality impact the relationship between ad viewability and actual purchase behavior to optimize the platform's viewability standards.

1
Analyze Viewability-Conversion Correlation
/about OpenPageRank IAB Categories Domain Ages
VIEWABILITY-CONVERSION CORRELATION ══════════════════════════════════════════════ VIEWABILITY IMPACT BY DOMAIN QUALITY: Premium domains (PageRank >7): Viewability: 78% | CVR: 3.8% | Correlation: 0.82 Mid-tier domains (PageRank 4-7): Viewability: 58% | CVR: 2.1% | Correlation: 0.67 Long tail (PageRank <4): Viewability: 34% | CVR: 0.8% | Correlation: 0.43 KEY INSIGHT: Viewability threshold of 50% is optimal cost-efficiency point Below 50%: CVR drops sharply (-67%) Above 70%: Diminishing returns on CVR improvement PageRank is strongest predictor of both viewability AND conversion
2
Set Domain-Specific Viewability Thresholds
Viewability Signal
Premium Domains (PageRank >7) — 78% Viewability — Strong correlation between PageRank and viewability confirmed: Premium publishers invest in good ad placement, resulting in 78% viewability and 3.8% CVR. These domains exceed the 50% threshold by wide margin. Bid premium justified: The 40% higher CPM is offset by 2.8x higher CVR, making premium inventory the most cost-efficient.
PREMIUM QUALITY — 78% viewability, 3.8% CVR, CPM premium justified
Viewability Signal
Long Tail (PageRank <4) — 34% Viewability — Below 50% optimal threshold. CVR drops to 0.8% on low-viewability inventory. PageRank is strongest predictor: domains under PageRank 4 rarely exceed 50% viewability. Recommendation: Apply viewability pre-filter for PageRank <4 domains, saving $5.6M/month in wasted spend on non-viewable impressions.
BELOW THRESHOLD — 34% viewability, 0.8% CVR, pre-filter recommended
3
Output: Viewability Standards Report

VIEWABILITY — CONVERSION ANALYSIS

VIEWABILITY STANDARD RECOMMENDATION: ══════════════════════════════════════════════ Optimal viewability threshold: 50% (cost-efficient) Revenue recovered from low-viewability filtering: $5.6M/quarter CVR improvement from viewability standards: +34% PageRank correlation with viewability: 0.82 (strongest predictor) Signals: PageRank, Domain Age, IAB Category, content quality

9Geographic Attribution Enhancement

AI agent enhances geographic attribution by using Countries enrichment data — determining which geographic audiences respond best to ad platform campaigns and optimizing ad spend allocation across markets based on domain-level geographic conversion intelligence.

1
Analyze Conversion Performance by Geography
/products Countries IAB Categories Personas
GEOGRAPHIC ATTRIBUTION — CONVERSION BY MARKET ══════════════════════════════════════════════ CAMPAIGN PERFORMANCE BY MARKET: US: CVR: 2.8% | ROAS: 6.8x | AOV: $134 | Budget share: 45% France: CVR: 2.3% | ROAS: 5.4x | AOV: $112 | Budget share: 15% UK: CVR: 2.4% | ROAS: 5.8x | AOV: $128 | Budget share: 12% Germany:CVR: 2.1% | ROAS: 4.8x | AOV: $118 | Budget share: 10% Japan: CVR: 1.9% | ROAS: 4.2x | AOV: $98 | Budget share: 8% Brazil: CVR: 1.4% | ROAS: 3.1x | AOV: $67 | Budget share: 5% GEO OPTIMIZATION RECOMMENDATION: Increase: UK (+3% budget share, under-invested vs ROAS) Decrease: Brazil (-2% budget share, low ROAS) Maintain: US, France (optimal allocation)
2
Identify Under/Over-Invested Geographic Markets
Geo Signal
UK — Under-Invested (ROAS 5.8x, Budget 12%) — UK delivers 3rd-highest ROAS (5.8x) but receives only 12% of budget (4th share). Domain intelligence shows UK publisher quality is high: avg PageRank 6.4, strong /legal compliance. Countries enrichment confirms 234M addressable profiles. Recommendation: Increase UK budget share by +3% from Brazil reallocation. Expected incremental revenue: $2.3M/quarter.
UNDER-INVESTED — UK ROAS 5.8x justifies +3% budget increase
Geo Signal
Brazil — Over-Invested (ROAS 3.1x, Budget 5%) — Brazil has lowest ROAS (3.1x) of top 6 markets. Domain intelligence diagnosis: 67% enrichment rate (lowest) limits targeting precision. Publisher quality lower (avg PageRank 3.8). Recommendation: Reduce Brazil budget by -2% and reallocate to UK/Germany until enrichment coverage improves to 85%+.
OVER-INVESTED — Brazil 3.1x ROAS, reduce -2% until enrichment improves
3
Output: Geographic Attribution Report

GEO ATTRIBUTION — OPTIMIZATION REPORT

GEOGRAPHIC OPTIMIZATION RESULTS: ══════════════════════════════════════════════ Markets analyzed: 34 countries Budget reallocation: UK +3%, Brazil -2%, others adjusted Revenue from geo optimization: +$8.9M/quarter ROAS improvement: +18% from geo-optimized allocation Signals: Countries, IAB Categories, Personas, advertiser domain geo

10Predictive ROAS Modeling

AI agent predicts campaign ROAS before launch by analyzing the target publisher domain mix and advertiser domain characteristics — using historical correlations between domain intelligence features and campaign outcomes to forecast performance with 87% accuracy.

1
Build ROAS Prediction from Domain Features
/products IAB Categories OpenPageRank Personas Countries Domain Ages
PREDICTIVE ROAS MODEL — DOMAIN FEATURES ══════════════════════════════════════════════ MODEL FEATURES (predicting campaign ROAS): Publisher PageRank distribution: Weight: 0.22 ■■■■■■■■■■■ IAB Category match score: Weight: 0.20 ■■■■■■■■■■ Persona alignment: Weight: 0.18 ■■■■■■■■■ Advertiser catalog depth (/products): Weight: 0.14 ■■■■■■■ Advertiser PageRank: Weight: 0.10 ■■■■■ Geographic match (Countries): Weight: 0.08 ■■■■ Domain Age (advertiser): Weight: 0.08 ■■■■ PREDICTION EXAMPLES: Nike on sports publishers: Predicted ROAS 7.8x (actual: 7.2x) Samsung on tech review sites: Predicted ROAS 6.4x (actual: 6.8x) Sephora on beauty/lifestyle: Predicted ROAS 8.2x (actual: 7.9x) MODEL ACCURACY: R² = 0.87 | Avg prediction error: 8.4%
2
Validate Predictions Against Historical Accuracy
Prediction Signal
Nike on Sports Publishers — Predicted 7.8x, Actual 7.2x — Prediction within 8.3% of actual (within acceptable range). Publisher PageRank distribution matched expectations. IAB Sports category match score was highest feature contributor. Persona alignment (Active Lifestyle) drove higher-than-expected conversion quality. Model performing well for sports vertical — increasing confidence weight for future predictions.
ACCURATE — Predicted 7.8x, actual 7.2x, within 8.3% error margin
Prediction Signal
Low-ROAS Campaign Prevented — Predicted 1.2x — Pre-launch analysis predicted 1.2x ROAS (below 2.0x threshold) for a fashion campaign targeting gaming sites. IAB category mismatch (Fashion vs Gaming: 18% match score) and Persona misalignment (Fashion Buyer vs Gamer: 12% overlap) drove low prediction. Campaign redirected to fashion/lifestyle publishers. Estimated $450K waste avoided.
CAMPAIGN BLOCKED — 1.2x predicted ROAS, $450K waste prevented
3
Output: Predictive ROAS Dashboard

PREDICTIVE ROAS — FORECASTING REPORT

PREDICTIVE ROAS MODEL RESULTS: ══════════════════════════════════════════════ Campaigns predicted: 4,567 this quarter Model accuracy: R² = 0.87 (87% of variance explained) Avg prediction error: 8.4% (within acceptable range) Low-ROAS campaigns prevented: 1,234 ($8.9M waste avoided) Revenue from prediction-optimized launches: +$12.3M/quarter Top features: PageRank, IAB match, Personas, catalog depth
Get in Touch

Interested in AI Agent Domain Intelligence?

For pricing, subscription options, custom database builds, or enterprise partnerships — contact us below.

Power Your AI Agents with Domain Intelligence

Subscribe to the AI Agent Domain Database — continuous access to 102M domains, 20 page types each, quarterly refreshes, and real-time change signals.

AI Agent Database View Pricing

Annual subscription includes quarterly data refreshes, change detection alerts, and priority API access.