Incrementality Testing for Auto Campaigns: GEO, PSA, and Ghost Ads

In today’s complex digital landscape, automotive marketers face a critical challenge: distinguishing which advertising efforts truly drive net-new vehicle sales versus those merely claiming credit for conversions that would have happened anyway. With significant dealer ad spend (see ROI metrics) and a majority flowing to digital channels, accurate measurement has become essential for maximizing ROI. Traditional attribution models, particularly last-click approaches, consistently overstate platform impact and fail to capture true causal relationships between ad exposure and vehicle purchases. Enter incrementality testing—the scientific method for determining your advertising’s actual lift on business outcomes.

At Demand Local, our LinkOne Data transforms incrementality testing from theoretical concept to practical reality by integrating first-party CRM and DMS data with precise campaign execution across all channels, enabling automotive marketers to measure what truly matters: incremental vehicle sales.

Key Takeaways

• Incrementality testing measures true causal impact of advertising by comparing exposed versus unexposed groups, unlike traditional attribution that only shows correlation
• GEO experiments divide markets into treatment and control regions, ideal for measuring regional campaign impact and foot traffic
• PSA testing uses public service announcements as placebo ads for control groups, providing precise measurement but at higher cost
• Ghost Ads methodology tracks “invisible” ads to control groups, offering high precision with potentially lower spend on supported platforms
• Proper implementation requires statistical planning, adequate sample sizes, and first-party data measurement tied to actual vehicle sales
• Each methodology has specific use cases: GEO for larger campaigns with geographic scale, PSA for precise measurement regardless of cost, Ghost Ads for continuous, cost-effective testing
• Demand Local’s platform enables precise execution of all three methodologies with real-time inventory sync and comprehensive attribution tracking

What Is Incrementality Testing and Why Auto Dealers Need It

Incrementality testing measures the causal, incremental impact of a marketing channel, campaign, or tactic on business results by comparing the behavior of users exposed to ads versus those who were not. In automotive marketing, this helps determine which advertising efforts actually drive net-new conversions like test drives, dealership visits, and vehicle sales, rather than just claiming credit for conversions that would have happened anyway through organic means.

Traditional attribution models used by digital ad platforms rely on inherently biased approaches, most often last-touch attribution, that overstate their impact. Platform-reported conversions often differ with site-side analytics or actual sales performance. As access to third-party data and user-level tracking diminishes, the accuracy of platform reporting is eroding further, making incrementality measurement critical for automotive marketers.

The Problem with Last-Click Attribution

Last-click attribution assigns 100% of conversion credit to the final touchpoint before purchase, ignoring all previous interactions that influenced the customer journey. This creates severe misallocation of budget in automotive marketing where:

• Car buyers typically visit multiple websites before purchase
• Customers interact with multiple touchpoints during their research process
• Brand awareness campaigns build consideration that may convert weeks later through a different channel
• Inventory-specific ads might receive last-click credit while broader awareness campaigns built the initial interest

Google has characterized incrementality testing as a gold standard for privacy-first impact measurement; these tests help marketers quantify the incremental revenue generated by a specific campaign—estimating the revenue your business would have missed without it.

Moving Beyond Correlation to Causation

The fundamental difference between traditional attribution and incrementality testing lies in their ability to establish causation versus mere correlation. When a customer sees your ad and later purchases a vehicle, attribution assumes the ad caused the purchase. However, that customer might have been planning to buy anyway—perhaps they were already in-market due to lease expiration or vehicle failure.

Incrementality testing solves this by creating randomized control groups that don’t see your ads, establishing a baseline of “what would have happened anyway.” The difference between your treatment group (exposed to ads) and control group (not exposed) represents your true incremental lift.

Measured notes that randomized experiments (test vs. control, exposed vs. non-exposed) are the way to infer the causal or incremental impact of media on sales.

GEO Experiments: Testing Regional Campaign Impact

Geographic experiments (GEO tests) divide markets into treatment regions (exposed to campaigns) and control regions (not exposed), measuring the difference in business outcomes between the two. This methodology works particularly well for automotive marketers because:

• Vehicle purchases are inherently local transactions
• Dealership networks often span multiple geographic markets
• Foot traffic and showroom visits can be measured through geofencing technology
• Regional market characteristics can be matched for valid comparisons

Setting Up Geographic Control Groups

Effective GEO experiments require careful market selection and matching. Key considerations include:

• Market matching: Select control markets with similar demographic profiles, economic conditions, and automotive purchasing patterns as treatment markets
• Geographic buffer zones: Create buffer areas between treatment and control markets to prevent spillover effects where consumers from control markets see ads intended for treatment markets
• Statistical power: Ensure sufficient population size in both treatment and control markets to detect meaningful differences
• Duration planning: Run tests long enough to capture full purchase cycles—often 30–90 days for automotive campaigns (a practical rule of thumb)

According to Haus Experimentation Platform, “Rather than relying on platform attribution or gut instinct, geo experiments use randomized, region-based testing to uncover the true causal impact of your campaigns. By comparing areas exposed to marketing against those that aren’t, you get a clearer picture of what actually drives results.”

Measuring Cross-Market Effects

One of the strengths of GEO experiments is their ability to measure effects beyond digital conversions. Automotive marketers can track:

• Foot traffic lift: Measured through geofencing technology and mobile location data
• Showroom visits: Verified through check-in systems or appointment scheduling
• Vehicle sales: Matched through VIN-specific sales data
• Service department visits: Tracked through customer service records

A DealerMarketing case study reported 140% relative lift (0.5% → 1.2%) in foot traffic, demonstrating the power of geographic testing to measure real-world business impact.

Demand Local’s LinkOne Data enables precise geo-targeted campaign execution with real-time inventory sync for accurate regional testing. By integrating DMS data with geographic targeting parameters, dealerships can measure exactly which vehicles moved as a result of regional campaigns, connecting ad spend directly to inventory turnover.

Public Service Announcements (PSA) as Placebo Ads

PSA testing serves real ads to both treatment and control groups but uses public service announcements for the control group instead of your actual marketing messages. This methodology maintains impression equality while isolating the impact of your specific creative and messaging.

Designing Effective PSA Controls

Effective PSA controls must match your actual ads in several key dimensions:

• Ad format: Same size, placement, and technical specifications
• User experience: Similar engagement patterns and interaction requirements
• Frequency: Equal impression frequency to avoid ad fatigue differences
• Platform: Served through the same ad networks and inventory sources

The goal is to ensure that the only difference between treatment and control groups is the actual marketing message—everything else remains identical.

Maintaining Test Validity

PSA testing provides highly precise measurement but can come with significant costs. Since you’re paying for impressions in both treatment and control groups, PSA tests can be expensive to execute at scale. However, they offer several advantages:

• Platform compatibility: Works within existing ad platform structures
• User-group precision: Can measure individual user behavior differences
• Creative isolation: Specifically tests the impact of your messaging and creative
• Behavioral insights: Reveals how your ads influence user actions beyond conversions

The primary limitation of PSA testing is cost—since you’re paying for control group impressions that don’t promote your brand, this methodology requires significant budget allocation. However, for high-value campaigns where precise measurement justifies the expense, PSA testing remains a gold standard for creative effectiveness measurement.

Ghost Ads: Invisible Testing for True Campaign Lift

Ghost Ads methodology represents a cost-effective approach to incrementality testing. Instead of serving actual ads to control groups, ghost ads track what would have been shown without actually serving the impression. This “bid-but-not-serve” approach records auction participation and targeting parameters while avoiding the cost of actual ad delivery to control groups.

How Ghost Bidding Works

The ghost ads process involves:

1. Auction participation: Your system participates in ad auctions for both treatment and control users
2. Impression logging: For treatment users, actual ads are served; for control users, the system logs what would have been shown
3. Synthetic control creation: The logged data creates a synthetic control group that mirrors actual auction conditions
4. Lift calculation: Compare conversion rates between actual treatment group and synthetic control group

Researchers report that, in their implementation context, advertisers can measure ad lift with similar precision while requiring substantially less spend.

Platform Requirements and Limitations

Ghost ads require sophisticated technical infrastructure and platform cooperation. Key requirements include:

• Real-time bidding integration: Ability to participate in auctions without serving ads
• Impression logging capabilities: Systems to record targeting parameters and auction outcomes
• User matching: Methods to connect ghost ad data with actual conversion data
• Statistical validation: Processes to ensure synthetic controls accurately represent real-world conditions

Feasibility varies by platform identity graphs and auction access; both prospecting and retargeting may be supported depending on implementation. The original work also notes implementations can record 100M+ predicted ghost ads per day.

Choosing the Right Testing Method for Your Dealership

Selecting the appropriate incrementality testing methodology depends on several factors specific to your dealership’s situation, campaign objectives, and available resources.

Small vs. Large Dealer Considerations

Small single-point dealerships should consider:

• GEO experiments if they operate in markets with sufficient geographic scale
• Ghost ads for cost-effective continuous measurement (if technical infrastructure exists)
• Starting with simpler methodologies and gradually building testing capabilities
• Partnering with platforms like Demand Local that provide testing infrastructure

Large multi-rooftop groups can leverage:

• Comprehensive GEO experiments across their regional footprint
• PSA testing for high-value campaigns where precise measurement justifies cost
• Integrated testing frameworks that combine multiple methodologies
• Dedicated testing budgets as part of overall marketing strategy

Single-Point vs. Multi-Rooftop Testing

Multi-rooftop dealer groups have unique advantages for incrementality testing:

• Natural market divisions: Different rooftops can serve as treatment and control groups
• Scale advantages: Larger customer bases provide better statistical power
• Resource sharing: Testing infrastructure and expertise can be centralized
• Cross-market validation: Results can be validated across similar markets

Single-point dealerships may need to:

• Partner with testing platforms that provide control group infrastructure
• Focus on specific campaign types rather than comprehensive testing
• Leverage industry benchmarks to supplement limited test data
• Start with inventory-specific tests that have clear success metrics

Budget requirements varies widely by scale & platform (estimate via pilot cost projection):

• GEO experiments: Moderate to high, depending on market size and duration
• PSA testing: High, due to dual impression costs
• Ghost ads: Low to moderate, primarily infrastructure costs

Setting Up Your Marketing Attribution Model

Incrementality testing works best when integrated into a comprehensive marketing attribution framework that connects campaign exposure to actual business outcomes. This requires robust first-party data infrastructure and clear conversion definitions.

First-Party Data Integration

Effective attribution models for automotive incrementality testing must include:

• CRM integration: Customer relationship management data linking ad exposure to actual purchases
• DMS connection: Dealer management system data providing VIN-specific sales information
• Inventory feeds: Real-time vehicle availability and pricing data
• Service records: Post-purchase engagement and loyalty metrics

Demand Local’s LinkOne Data integrates CRM/DMS data for comprehensive attribution tracking across all customer touchpoints, creating the foundation for accurate incrementality measurement.

Building Your Attribution Framework

A robust attribution framework includes:

• Clear conversion definitions: What constitutes a successful outcome (VDP view, lead, showroom visit, sale)
• Multi-touch attribution windows: Time periods for crediting ad exposure to conversions
• Cross-device tracking: Methods to connect user behavior across devices
• Offline conversion tracking: Systems to link digital exposure to in-store purchases

The framework should be designed to support incrementality testing by providing clean separation between treatment and control group outcomes while maintaining data integrity across all touchpoints.

Implementing Marketing Attribution Software for Testing

Successful incrementality testing requires specialized software infrastructure that can execute tests, collect data, and provide actionable insights. Key capabilities include:

Essential Software Features

• Randomized group assignment: Automated systems for creating statistically valid test and control groups
• Real-time data collection: APIs that capture impression, click, and conversion data as it happens
• Statistical analysis tools: Built-in calculators for confidence intervals, p-values, and statistical significance
• Dashboard visualization: User-friendly interfaces for monitoring test progress and results
• Automated reporting: Scheduled reports that deliver insights without manual intervention

Integration with DMS Systems

Automotive-specific attribution software must integrate seamlessly with common DMS platforms including:

• CDK Global
• Reynolds and Reynolds
• Tekion
• Dealertrack
• VinSolutions

These integrations enable VIN-specific matchback analysis, connecting specific ad campaigns to actual vehicle sales and providing definitive ROI calculations.

Demand Local’s platform provides pre-built integrations that minimize manual file uploads and enable rapid campaign launch, with many clients launching initial campaigns within about a week (Inventory Marketing).

Statistical Significance and Sample Size for Auto Campaigns

Proper statistical planning is essential for valid incrementality testing results. Automotive campaigns present unique challenges due to extended purchase cycles and relatively low conversion rates.

Calculating Required Test Duration

Key statistical considerations include:

• Minimum detectable effect: The smallest lift you want to be able to detect
• Statistical power: Typically 80% power to detect the minimum effect size
• Confidence level: Usually 95% confidence (p < 0.05)
• Baseline conversion rate: Historical conversion rates for similar campaigns
• Sample size requirements: Number of users needed in each group

Allocate control vs. treatment based on a power analysis that considers baseline conversion rate, minimum detectable effect, confidence level, and desired power. Larger control groups often provide more reliable results.

Avoiding Common Statistical Errors

Common pitfalls in automotive incrementality testing include:

• Insufficient sample sizes: Leading to inconclusive or misleading results
• Premature test termination: Stopping tests before statistical significance is achieved
• Multiple hypothesis testing: Running too many tests simultaneously without proper adjustment
• Ignoring seasonality: Failing to account for natural fluctuations in automotive purchasing patterns
• Campaign interference: Allowing treatment and control groups to influence each other

Automotive campaigns typically require longer test periods than other industries due to extended consideration cycles. While e-commerce tests might run for 7–14 days, automotive incrementality tests often need 30–90 days to capture meaningful results (a practical rule of thumb).

Measuring Inventory Turn and Sales Lift

The ultimate measure of automotive advertising effectiveness is its impact on inventory movement and vehicle sales. Incrementality testing provides the causal link between ad spend and these business outcomes.

Tracking Aged Inventory Movement

Aged inventory represents significant financial risk for dealerships, making it a critical focus for incrementality testing. Key metrics include:

• Days-on-lot reduction: Measuring how quickly specific vehicles sell after campaign exposure
• Aged unit sales lift: Comparing sales of vehicles over 60/90/120 days old between treatment and control groups
• Gross profit impact: Calculating the financial benefit of moving aged inventory versus holding costs
• Model-specific performance: Identifying which vehicle models respond best to specific advertising approaches

Demand Local’s Inventory Marketing solution enables dynamic VIN-level ads with real-time updates, allowing precise measurement of inventory movement from specific campaigns. Case studies show aged EV units selling over about 90 days via dynamic display + CTV.

Connecting Ad Spend to Vehicle Sales

True incrementality measurement requires connecting digital ad exposure to actual vehicle purchases. This involves:

• VIN-specific matchback: Linking ad exposure data to specific vehicle sales records
• Customer identity resolution: Connecting anonymous ad exposure to known customer profiles
• Multi-touch attribution: Understanding how multiple ad exposures contribute to final purchase decisions
• Lifetime value calculation: Measuring not just immediate sales but long-term customer value

Key automotive-specific metrics include:

• VDP views: Vehicle detail page views as leading indicators of purchase intent
• Lead quality scores: Measuring the likelihood that leads will convert to actual sales
• Showroom visit attribution: Connecting digital exposure to in-person dealership visits
• Service department tracking: Monitoring post-purchase engagement and loyalty

Common Pitfalls in Automotive Incrementality Testing

Even well-designed incrementality tests can fail due to implementation errors and external factors. Understanding these pitfalls is essential for valid results.

Accounting for External Factors

Automotive markets are influenced by numerous external factors that can contaminate test results:

• OEM incentives: Manufacturer rebates and special financing offers can override advertising effects
• Seasonality: Natural fluctuations in automotive purchasing patterns throughout the year
• Competitive actions: Rival dealerships’ promotional activities can influence results
• Economic conditions: Broader economic trends affect consumer purchasing behavior
• Inventory availability: Stock levels and vehicle availability can limit conversion opportunities

Effective testing requires either controlling for these factors or ensuring they affect treatment and control groups equally.

Maintaining Test Integrity

Common implementation errors include:

• Selection bias: Non-random assignment of users to treatment and control groups
• Contamination: Treatment and control groups influencing each other (spillover effects)
• Promotional overlap: Other marketing activities interfering with test campaigns
• Measurement errors: Inaccurate conversion tracking or data collection issues
• Platform changes: Ad platform algorithm updates affecting test conditions

Ghost ads methodology helps minimize some of these issues by using synthetic controls that are less susceptible to external contamination, while GEO experiments provide natural isolation through geographic separation.

How Demand Local Powers Incrementality Testing Success

While incrementality testing concepts are straightforward, successful implementation requires sophisticated technology infrastructure and automotive industry expertise. Demand Local’s platform delivers the capabilities automotive marketers need to execute valid, actionable incrementality tests across all methodologies.

Our solutions excel where others fall short through:

• First-Party Data Integration: LinkOne Data seamlessly connects CRM, DMS, and inventory feeds, creating the clean data foundation essential for valid incrementality measurement
• Precise Campaign Execution: Real-time inventory sync ensures treatment groups see accurate, available vehicles while control groups receive appropriate alternatives (PSA or ghost ads)
• Multi-Channel Measurement: Comprehensive tracking across search, social, CTV, and DOOH enables holistic incrementality testing across the entire customer journey
• Automotive-Specific Attribution: VIN-level matchback and dealership-specific tracking provide definitive connection between ad exposure and vehicle sales
• Rapid Implementation: Pre-built integrations with major DMS platforms enable campaign launch within days (Inventory Marketing), not weeks or months

Demand Local’s platform has driven strong results for dealerships implementing incrementality testing, with dealers reporting reductions in cost-per-lead after integrating vehicle listing ads with SEM and comprehensive attribution tracking.

Our automotive expertise ensures that incrementality tests are designed with industry realities in mind—including extended purchase cycles, OEM compliance requirements, and the unique dynamics of new versus used vehicle marketing. Whether you’re executing GEO experiments across multiple markets, PSA testing for creative validation, or ghost ad methodology for continuous optimization, Demand Local provides the infrastructure and expertise to make incrementality testing work for your dealership.

FAQs on Incrementality Testing for Automotive Campaigns

Q: What’s the minimum budget needed for incrementality testing in automotive?

A: Most incrementality tests require sufficient scale to achieve statistical significance. For GEO experiments, a rule of thumb is sufficient population per power analysis (commonly >100k, but depends on baseline and MDE). PSA testing costs vary by platform and generally require larger allocations due to impressions delivered to both groups. Ghost ads are often more cost-effective where supported, with infrastructure costs being the primary investment. Many dealerships start with smaller-scale tests focused on specific inventory segments (like aged units) before expanding.

Q: How long should I run a geo-experiment for reliable results?

A: Automotive geo-experiments often require 30–90 days due to extended purchase cycles. Unlike e-commerce with immediate conversions, car buyers may take weeks or months to complete their purchase after initial ad exposure. Exact duration depends on baseline conversion rate, desired statistical power, and the effect size you aim to detect.

Q: Can I test incrementality on both new and used inventory simultaneously?

A: Yes, but it’s usually more effective to test them separately due to different cycles, motivations, and competitive dynamics. If testing both simultaneously, ensure your model accounts for these differences or create separate treatment and control groups for each.

Q: Which platforms support ghost ad testing for auto dealers?

A: Native Ghost Ads are not universally available. Major platforms offer their own lift measurement products, and third-party measurement solutions can create synthetic controls by integrating with large ad networks. Availability and mechanics vary by platform and access.

Q: How do I account for OEM co-op spending in incrementality tests?

A: Co-op requirements can complicate testing. Strategies include: testing within OEM-compliant creative parameters (e.g., PSA alternatives), coordinating with your OEM representative for methodology approvals, focusing on non-co-op eligible campaigns first, or using ghost ad approaches that don’t require serving control-group ads—while maintaining OEM compliance.

Q: What’s the difference between incrementality testing and A/B testing?

A: A/B testing compares two or more versions promoting your brand to determine the better performer. Incrementality testing compares a treatment group exposed to your ads versus a control group that isn’t (or sees placebo ads) to measure true incremental lift above what would have happened organically.