Make confident, data-driven decisions with actionable ad spend insights.
© 2026 DataCops. All rights reserved.
14 min read
Moving from optimizing for simple 'Conversions' to optimizing for 'Conversion Value' is the single most effective lever available to modern performance marketers. However, the move is often hampered by the same underlying data integrity issues that plague standard conversion bidding. Value-Based Bidding (VBB) requires high-fidelity, high-volume data to succeed.
Orla Gallagher
PPC & Paid Social Expert
Last Updated
December 15, 2025
The Problem: Your Google Ads Maximize Conversion Value campaign optimizes toward revenue. Algorithm learns average order value (AOV) is $800 from tracked conversions. Bids conservatively on high-value customer segments. Backend data shows true AOV is $1,200 but ad blockers prevented tracking 35% of high-value purchases. Algorithm systematically underbids on most profitable customers by 33% because never sees their true value.
The Reason: Ad blockers prevent conversion value tracking for 25-35% of users, disproportionately affecting high-value customers (privacy-conscious, use Safari/Brave). Algorithm trains on incomplete sample showing false $800 AOV when true AOV is $1,200. Bot traffic creates fake low-value conversions ($50 average) further deflating perceived customer value. VBB bids based on corrupted $700 blended average (mix of true $800 visible + $50 bot pollution) instead of actual $1,200, losing high-value auctions.
The Solution: Implement first-party conversion tracking via CNAME capturing 95%+ of conversions including high-value purchases. Filter bot traffic creating fake low-value signals. Algorithm sees true $1,200 AOV from complete, clean data. VBB increases bids 40-50% on profitable customer segments. Wins high-value auctions previously lost, improves ROAS 40-60% by optimizing toward actual customer profitability not corrupted sample.
Value-Based Bidding (VBB) optimizes campaigns toward conversion value (revenue or profit) instead of conversion count, maximizing total return not just conversion volume.
Target CPA (conversion count):
Optimizes: Number of conversions.
Goal: Get 100 conversions at $50 each.
Treats: All conversions equally valuable.
Problem: $100 sale same priority as $1,000 sale.
Value-Based Bidding (conversion value):
Optimizes: Total conversion value.
Goal: Maximize revenue/profit from spend.
Treats: High-value conversions more important.
Result: $1,000 sale gets 10x higher bid than $100 sale.
VBB bidding strategies:
Maximize Conversion Value: Spend entire budget to generate maximum total revenue.
Target ROAS: Achieve specific return on ad spend (e.g., 400% = $4 revenue per $1 spent).
Value calculation example:
Campaign budget: $10,000
Target CPA approach:
200 conversions at $50 each
Mix: 180 low-value ($100) + 20 high-value ($1,000)
Total revenue: $38,000 (380% ROAS)
VBB approach:
150 conversions (fewer total)
Mix: 100 low-value ($100) + 50 high-value ($1,000)
Total revenue: $60,000 (600% ROAS)
VBB prioritizes profitable customers over volume.
VBB algorithm learns customer value patterns from conversion data, but ad blockers hide 25-35% of high-value conversions, causing systematic underbidding on most profitable segments.
VBB learning process:
Analyzes: Which user signals correlate with high conversion value.
Learns: iPhone users + premium brand searches = $1,200 AOV.
Optimizes: Bid $30 for these high-value patterns.
Incomplete data problem:
100 actual high-value conversions ($1,200 AOV).
Ad blockers prevent tracking for 35 conversions.
Algorithm sees only 65 conversions.
Calculates: $1,200 × 65 = $78,000 total value.
Observed AOV: $78,000 ÷ 100 users = $780 (not true $1,200).
Bidding consequence:
Algorithm thinks high-value segment worth $780.
Bids: $20 per click ($780 AOV × 2.5% conversion rate).
Actual value: $1,200 (54% higher).
Should bid: $30 per click.
Loses auctions: Competitor bids $28, you bid $20.
Miss profitable opportunities by underbidding.
Scale of problem:
25-35% of high-value conversions hidden.
Algorithm undervalues profitable segments by 30-50%.
Budget wasted on low-value traffic algorithm can see.
High-value traffic under-served because appears less valuable.
Ad blockers disproportionately affect high-value customers (privacy-conscious users, executives, technical buyers), hiding their conversion value from VBB algorithm.
Customer privacy correlation:
Low-value customers ($50-$200): 15-20% use ad blockers.
Medium-value customers ($200-$500): 25-30% use ad blockers.
High-value customers ($500-$2,000): 35-45% use ad blockers (privacy-conscious, use Safari/Brave).
Impact on value visibility:
200 total conversions:
100 low-value ($150 average): 18 blocked (82 tracked)
60 medium-value ($300 average): 17 blocked (43 tracked)
40 high-value ($1,200 average): 16 blocked (24 tracked)
What algorithm sees:
Low-value: 82 tracked × $150 = $12,300
Medium-value: 43 tracked × $300 = $12,900
High-value: 24 tracked × $1,200 = $28,800
Total: 149 tracked conversions, $54,000 value
Calculated AOV: $54,000 ÷ 149 = $362
Reality (complete data):
Low-value: 100 × $150 = $15,000
Medium-value: 60 × $300 = $18,000
High-value: 40 × $1,200 = $48,000
Total: 200 conversions, $81,000 value
True AOV: $81,000 ÷ 200 = $405
Bidding distortion:
Algorithm bids for $362 AOV (observed).
Should bid for $405 AOV (actual).
Underbids by 12% on average, 40%+ on high-value segments.
Loses most profitable customers to competitors.
Bot traffic creates fake low-value conversions deflating perceived customer value, causing VBB to underbid on real high-value users.
Bot conversion patterns:
Bots complete forms instantly (fake leads).
Bots add cheap items to cart (fake low-value purchases).
Bots never convert to high-value purchases.
Impact on value calculations:
Real human conversions:
150 conversions
Average value: $600
Total: $90,000
Bot conversions:
30 fake conversions
Average value: $50 (low-value items)
Total: $1,500
What algorithm sees:
Total: 180 conversions (150 human + 30 bot)
Total value: $91,500
Calculated AOV: $91,500 ÷ 180 = $508
True human AOV: $90,000 ÷ 150 = $600
Algorithm undervalues humans by 15%.
Bidding consequence:
VBB bids for $508 AOV (bot-polluted).
Should bid for $600 AOV (human-only).
Underbids 15%, loses high-value human auctions.
Budget partially wasted on bot patterns.
Element Third-Party Tracking (Incomplete) First-Party Tracking (Complete)
High-value conversions (40 actual) 24 tracked (16 blocked, 40% loss) 38 tracked (2 natural loss, 5% loss)
Medium-value conversions (60 actual) 43 tracked (17 blocked, 28% loss) 57 tracked (3 natural loss, 5% loss)
Low-value conversions (100 actual) 82 tracked (18 blocked, 18% loss) 95 tracked (5 natural loss, 5% loss)
Bot conversions included 30 fake low-value ($50 avg) 0 (filtered before tracking)
Total tracked conversions 179 (149 human + 30 bot) 190 (bot-free)
Total tracked value $56,500 (bot-polluted) $79,900 (complete)
Algorithm observed AOV $316 (deflated) $421 (accurate)
True AOV $405 $405
Bidding accuracy Underbids 22% Accurate
High-value segment bid $15 (too low) $25 (optimal)
ROAS achieved 380% (underbidding on profitable) 580% (optimal allocation)
Static value (flawed VBB):
All conversions assigned same value.
Example: Every lead worth $50.
Algorithm cannot differentiate high-value from low-value.
Bids equally for executive demo request and student inquiry.
Dynamic value (true VBB):
Conversions assigned actual revenue or predicted value.
E-commerce: Order value $89 or $1,247.
B2B: Enterprise lead $5,000, SMB lead $500.
Algorithm learns which patterns drive high vs low value.
Dynamic value implementation:
E-commerce:
Pass actual order value: $89, $234, $1,567.
Algorithm learns: Premium brand searches = $800+ orders.
Bids highest for high-AOV indicators.
B2B lead gen:
Form submit: Capture company size, industry.
Assign predicted value: Fortune 500 = $10,000, SMB = $500.
Algorithm optimizes toward high-value company profiles.
Offline conversion import:
Lead converts to sale 6 months later: $50,000 deal.
Send actual deal value back to platform.
Algorithm learns initial ad click led to $50,000 outcome.
Increases bids for similar future patterns.
Week 1-2: Establish value tracking
E-commerce: Pass actual transaction value in conversion tracking.
B2B: Assign predicted values based on lead quality signals.
Verify: Backend order value matches reported conversion value.
Gap >15% indicates tracking issues.
Week 3-4: Implement first-party tracking
Deploy analytics via CNAME (analytics.yourstore.com).
Verify 95%+ conversion value capture (not just count).
Compare: High-value order tracking before (65%) vs after (95%).
Week 5-6: Enable bot filtering
Activate bot detection filtering low-value fake conversions.
Remove 15-25% pollution from value calculations.
Verify: Average order value increases 10-20% post-filtering.
Week 7-8: Baseline establishment
Run Maximize Conversion Value for 30 days.
Algorithm learns from complete, clean value data.
Document: True AOV, ROAS, value distribution.
Week 9: Set Target ROAS
Calculate: Current ROAS from complete data.
Example: $10,000 spend, $50,000 revenue = 500% ROAS.
Set initial target: 450% (10% buffer below observed).
Week 10-11: Learning phase
Algorithm optimizes toward 450% ROAS target.
7-14 day learning with complete value signals.
Monitor: Maintains budget spend, shifts toward high-value.
Week 12+: Gradual optimization
Increase Target ROAS by 25-50 points every 2 weeks.
Example: 450% → 475% → 500% → 525%.
Stop when volume decreases significantly.
Mistake 1: Using static values for all conversions
Assign all leads $100 value regardless of quality.
Algorithm cannot differentiate $50,000 enterprise opportunity from $500 SMB lead.
Bids equally for all, wastes budget on low-value.
Fix: Implement dynamic value based on lead attributes or actual deal size.
Mistake 2: Not tracking value for high-value conversions
Ad blockers prevent value tracking for 35% of $1,000+ orders.
Algorithm sees mostly low-value $100-$300 orders.
Underestimates high-value customer segment worth.
Fix: First-party tracking captures 95%+ including high-value.
Mistake 3: Including bot conversions in value calculations
Bots create 30 fake $50 conversions monthly.
Deflates average value from $600 to $500.
Algorithm underbids on real humans by 17%.
Fix: Filter bots before sending value data to algorithm.
Mistake 4: Setting Target ROAS too high initially
True ROAS 500%, set target 600% immediately.
Algorithm restricts spending (target appears unachievable).
Budget underutilized, miss profitable opportunities.
Fix: Start at 450% (below current 500%), increase gradually.
Mistake 5: Not using offline conversion import
B2B lead converts to $50,000 sale 6 months later.
Never report final value back to platform.
Algorithm thinks lead worth $100 (form submit value).
Fix: Send actual deal value when closes, update algorithm learning.
Check 1: Value data completeness
[ ] Backend high-value orders (>$500): _____ monthly
[ ] Platform tracked high-value orders: _____ monthly
[ ] Gap: _____%
[ ] If gap >25%, losing most profitable conversion data
Check 2: Average order value accuracy
[ ] Backend calculated AOV: $_____
[ ] Platform reported AOV: $_____
[ ] Difference: _____%
[ ] If platform AOV 15%+ lower, tracking incomplete
Check 3: Value distribution
[ ] Platform shows value range: $50-$2,000
[ ] Backend shows value range: $50-$2,000
[ ] If platform missing high-end, ad blockers hiding premium customers
Check 4: Bot pollution check
[ ] Review conversions for $10-$50 low-value patterns
[ ] Check for instant checkout, zero engagement time
[ ] Estimate bot %: Should be <5% after filtering
Check 5: Static vs dynamic values
[ ] Check if all conversions assigned same value
[ ] If yes, not true VBB (value-adjusted CPA only)
[ ] Implement dynamic value assignment
What is Value-Based Bidding?
Value-Based Bidding (VBB) optimizes campaigns toward conversion value (revenue or profit) instead of conversion count. Algorithm bids higher for users likely to generate high-value purchases ($1,000+) and lower for low-value purchases ($100), maximizing total return not conversion volume. Strategies include Maximize Conversion Value and Target ROAS.
Why does Value-Based Bidding fail with incomplete data?
VBB fails when ad blockers hide 25-35% of high-value conversions (privacy-conscious customers use Safari, Brave). Algorithm sees false $800 AOV when true AOV is $1,200 (35% blocked). Bids 33% too low for profitable segments, loses high-value auctions to competitors. Budget wasted on low-value traffic algorithm can fully observe.
How do ad blockers affect Value-Based Bidding?
Ad blockers disproportionately affect high-value customers (35-45% use blockers) vs low-value customers (15-20% use blockers). Algorithm sees mostly low-value conversions that bypass blockers, underestimates profitable customer segment worth by 30-50%. Systematic underbidding on best customers causes 40-60% ROAS loss.
What is the difference between static and dynamic values?
Static value assigns same worth to all conversions (all leads $100), cannot differentiate high from low value. Dynamic value assigns actual revenue ($89, $450, $1,234) or predicted profit (enterprise lead $5,000, SMB $500), enabling algorithm to learn profitable patterns and bid accordingly. Only dynamic enables true value-based optimization.
How do I set Target ROAS for Value-Based Bidding?
Calculate current ROAS from complete conversion value data (not incomplete platform reporting): Total Revenue ÷ Total Spend × 100. If current 500% ROAS with complete data, set initial target 450% (10% buffer). Gradually increase 25-50 points every 2 weeks: 450% → 475% → 500% → 525% until volume decreases.
What is offline conversion import for VBB?
Offline conversion import sends actual deal value back to platform when B2B leads close months after initial conversion. User submits lead form (assigned $100), converts to $50,000 sale 6 months later. Send $50,000 actual value to platform, algorithm learns initial ad click drove high-value outcome, increases bids for similar patterns in future.
DataCops provides first-party analytics platform that captures 95%+ of conversion value data including high-value purchases, enabling Value-Based Bidding to optimize on true customer profitability not corrupted sample.
Complete high-value conversion tracking:
First-party script from analytics.yourstore.com bypasses ad blockers.
Captures 95%+ of high-value conversions ($500-$2,000+).
Standard third-party tracking captures 65% (35% blocked).
Algorithm sees true customer value distribution.
Accurate AOV calculation:
Before first-party (incomplete):
High-value conversions: 24 of 40 tracked (40% blocked)
Calculated AOV: $362
True AOV: $405
11% undervaluation
After first-party (complete):
High-value conversions: 38 of 40 tracked (5% natural loss)
Calculated AOV: $421
True AOV: $405
Accurate representation
Bot-filtered value data:
Removes fake low-value bot conversions ($20-$50).
Bots create 15-20% of reported low-value purchases.
Exclusion prevents AOV deflation.
Algorithm optimizes on human value patterns only.
VBB bidding accuracy:
Incomplete data: Algorithm bids for false $316 AOV (30% too low).
Complete data: Algorithm bids for true $421 AOV (optimal).
High-value segment: Bids increase from $15 to $25 (67% correction).
Wins profitable auctions previously lost to competitors.
Dynamic value assignment support:
E-commerce: Actual transaction values passed automatically.
B2B: Predictive lead scoring based on form data:
Company size: Fortune 500 = $10,000, SMB = $500
Industry: Enterprise software = $8,000, Retail = $1,500
Job title: C-level = $7,500, Manager = $2,000
Offline conversion integration:
Syncs with Salesforce, HubSpot for closed deal values.
Lead converts 6 months later: Send actual $50,000 to platform.
Algorithm learns long-term value, not just initial $100 lead value.
Optimizes for deals not just form fills.
Target ROAS optimization:
Complete data reveals true 500% ROAS vs false 380%.
Recommends initial target: 450% (achievable with buffer).
Gradual increase schedule: 450% → 475% → 500% → 525%.
Prevents premature spend restrictions from overaggressive targets.
Cross-platform value consistency:
Same complete value data sent to Google Ads and Meta.
Both platforms optimize on accurate customer profitability.
Eliminates contradictory ROAS reporting across channels.
ROAS improvement:
Before (incomplete value data):
Observed AOV: $316 (deflated)
ROAS: 380% (underbidding on profitable)
After (complete value data):
Observed AOV: $421 (accurate)
ROAS: 580% (optimal allocation to high-value)
52% ROAS improvement from data quality alone.
Implementation timeline:
Week 1-2: CNAME DNS setup, first-party script deployment
Week 3-4: Value tracking verification (95%+ high-value capture)
Week 5-6: Bot filtering calibration, AOV correction
Week 7-8: Maximize Conversion Value baseline with complete data
Week 9-10: Target ROAS implementation with accurate starting point
Week 11-12: Gradual optimization, 40-60% ROAS improvement
Platform automatically captures complete conversion value data including blocked high-value purchases and syncs to VBB algorithms for optimal profit-based bidding with no manual work required.
Key Takeaways:
Value-Based Bidding optimizes toward conversion value not count, but ad blockers hide 25-35% of high-value conversions causing systematic underbidding
High-value customers (35-45% use ad blockers) get blocked more than low-value customers (15-20%), algorithm undervalues profitable segments by 30-50%
Algorithm sees false $800 AOV (incomplete data) when true AOV is $1,200, underbids 33% on best customers, loses high-value auctions
Bot traffic creates fake low-value conversions ($20-$50) deflating AOV 10-20%, causing VBB to underbid on real high-value humans
First-party tracking via CNAME captures 95%+ of high-value conversions instead of 65%, reveals true customer value for accurate bidding
Use dynamic value assignment (actual revenue or predicted LTV) not static values ($100 per lead) to enable true value-based optimization
Set initial Target ROAS 10% below current ROAS with complete data, increase gradually by 25-50 points every 2 weeks
Complete value data improves ROAS 40-60% by enabling algorithm to correctly identify and bid aggressively on most profitable customer segments
