TL;DR
A geo-lift test is an incrementality experiment that uses geographic markets as the test and control groups — run campaigns in some markets, withhold them in others, then compare conversion rates. It's the practical alternative to platform holdout tools for measuring true ad impact. Typical B2B SaaS geo-lift tests run 4–6 weeks across matched market pairs.
What is a geo-lift test?
A geo-lift test (also called geographic holdout, geo experiment, or geo incrementality test) is an A/B experiment that uses geographic markets — cities, DMAs, or countries — as the test and control units instead of individual users. Ads run in the test markets; the control markets see no ads. The difference in conversion rate between the two groups measures incremental lift.
Geo-lift tests exist because most ad platforms can't randomise at the user level without significant data-sharing. For channels like linear TV, out-of-home, podcast ads, or any cross-device campaign, user-level holdouts are technically infeasible. Geographic experiments sidestep the technical constraint by randomising at the market level instead.
For B2B SaaS operators, geo-lift tests are most commonly used to measure the incremental impact of brand-awareness campaigns (LinkedIn, podcast, display) that lack last-click attribution. For D2C brands, they're used to validate whether paid media is driving incremental purchases or simply claiming credit for organic demand.
Why geo-lift tests matter for operators
Without incrementality measurement, operators scale based on attributed ROAS — a metric that overstates ad impact for any channel where organic demand exists. Geo-lift tests break the correlation-causation problem by showing what actually happens to conversion rate when ads are present versus absent.
The practical implication is budget reallocation. A brand running $80,000/month on awareness channels and seeing 2.8× attributed ROAS might run a geo-lift test and discover true incremental ROAS of 0.9× — the ads are claiming credit for conversions that would have happened organically. Redirecting that $80K to higher-lift channels produces the same total revenue at lower cost.
A typical D2C brand running a geo-lift test for the first time discovers that 30–50% of attributed revenue from retargeting campaigns is organic — customers who would have repurchased without seeing the retargeting ad. This is one of the most common profit leaks in D2C marketing.
How to run a geo-lift test
Requirements for a valid geo-lift test:
Geo-Lift Test Structure:
1. Select matched market pairs
— Match on: historical conversion rate, population size,
income level, product penetration
— Minimum 3 test markets, 3 control markets for significance
— Tools: Google's Market Finder, Nielsen DMA data
2. Define test period
— B2B SaaS: 4–6 weeks minimum (longer sales cycles)
— D2C: 2–4 weeks (faster conversion window)
— Must cover at least 1 full conversion cycle
3. Run campaign in test markets only
— Strictly exclude control markets from all targeting
— Use IP targeting, geographic bid modifiers at market level
4. Measure lift
Lift = ((Test Conversion Rate − Control Conversion Rate)
/ Control Conversion Rate) × 100
Example:
Test markets (6 cities with ads): 0.31% conversion rate
Control markets (6 cities, no ads): 0.22% conversion rate
Lift = ((0.31 − 0.22) / 0.22) × 100 = 40.9% - Minimum 6 matched market pairs for statistical significance at 80% power
- Pre-experiment period of equal length to check balance before running
- Clean campaign execution — zero bleed-over into control markets
- Consistent measurement methodology (conversions defined identically in both groups)
- No concurrent campaigns or promotions that could confound the result
Geo-lift test benchmarks and outcomes
| Campaign type | Typical lift range | Typical test duration | Common finding |
|---|---|---|---|
| B2B SaaS — LinkedIn brand awareness | 10–30% | 4–6 weeks | 20–40% of attributed pipeline is organic |
| B2B SaaS — podcast / audio ads | 5–20% | 6–8 weeks | Long delay before lift shows; hard to measure short-term |
| D2C — prospecting (Meta/Google) | 25–50% | 2–4 weeks | Strong lift; prospecting is genuinely incremental |
| D2C — retargeting (Meta/Google) | 5–20% | 2–3 weeks | Low lift; most would have repurchased organically |
| D2C — TV / CTV / OOH | 10–35% | 4–6 weeks | Lift appears with 2–3 week delay post-exposure |
Sources: Google Geo Experiments methodology docs; Nielsen Attribution benchmarks 2024; Meta Lift Studies 2024; Fairview customer data.
Common mistakes when running geo-lift tests
1. Using markets that are too small for statistical significance. A geo-lift test with 3 test cities and 3 control cities typically doesn't generate enough conversions to reach 95% confidence. Use power analysis before running — calculate the sample size needed given your current conversion rate and expected lift.
2. Selecting poorly matched markets. Test and control markets must be demographically and historically similar. Matching New York (test) against rural Montana (control) will produce noise, not signal. Use 6–12 months of pre-experiment conversion data to verify baseline similarity before running.
3. Running the test too short. B2B SaaS sales cycles are 2–12 weeks. A 2-week geo-lift test on a B2B product measures top-of-funnel response, not closed revenue. Run the test long enough for conversions to complete the full funnel from the campaigns in the test period.
4. Not controlling for seasonality. Running a geo-lift test over a holiday period, product launch, or major news event in one market group introduces confounding variance. Choose test windows that are neutral for both groups.
5. Treating the lift number as permanent. Lift from a specific campaign, audience, and message is point-in-time. Creative fatigue, competitive changes, and audience saturation reduce lift over time. Re-run geo-lift tests at least quarterly on your largest channels.
How Fairview connects geo-lift results to margin
Fairview's Margin Intelligence module connects ad-platform spend to CRM and revenue data so operators can apply geo-lift results to recalibrate channel-level ROAS. When a geo-lift test reveals that retargeting lift is 12% rather than the attributed 4× ROAS, Fairview adjusts the channel's efficiency score to reflect true incremental return.
The Next-Best Action Engine flags attribution-inflation signals: "Retargeting ROAS held flat for 30 days while organic direct traffic increased 22%. Attribution inflation risk: a geo-lift test is recommended before scaling retargeting spend further."
Companies using Fairview that run geo-lift tests and feed results into the platform typically reallocate 20–30% of their paid budget within two quarters to higher-lift channels.
At a glance
- Category
- Marketing Metrics
- Related
- 5 terms
Frequently asked questions
What is a geo-lift test in simple terms?
An experiment where you run ads in some cities or regions and hold back ads in a matched set of similar cities, then compare conversion rates between the two groups. The difference is the lift your ads are actually causing — not just correlating with.
How is a geo-lift test different from a holdout test?
A holdout test withholds ads from a random subset of users on a single platform. A geo-lift test withholds ads from entire geographic markets. Holdout tests are better when the platform supports user-level randomisation (Meta, Google). Geo-lift tests are better for channels that can't do user-level holdouts — TV, podcast, out-of-home, cross-device campaigns.
How many markets do you need for a valid geo-lift test?
A minimum of 6 matched market pairs (3 test, 3 control). For 80% statistical power and a 20% lift, you'll typically need 8–12 pairs depending on your baseline conversion rate. Use a power analysis calculator to determine the right number before running — under-powered tests produce inconclusive results.
How long should a geo-lift test run?
D2C / e-commerce: 2–4 weeks. B2B SaaS: 4–8 weeks minimum. The test must cover at least one full conversion cycle from ad exposure to closed deal (or purchase for D2C). Running a 2-week test on a product with a 6-week sales cycle measures only top-of-funnel response, not revenue impact.
What is a normal geo-lift test result?
For D2C prospecting campaigns: 25–50% lift is typical; these are genuinely incremental. For D2C retargeting: 5–20%, because most retargeted customers would have repurchased anyway. For B2B awareness: 10–30%. Results below 5% suggest the channel is primarily claiming credit for organic demand rather than generating incremental revenue.
Sources
- OpenView SaaS Benchmarks 2025
- Pavilion Operator Survey 2024
- Common Thread Collective D2C Benchmarks 2025
- ProfitWell Research
- Fairview customer data (B2B SaaS + D2C, 2025)
Fairview is an operating intelligence platform that connects geo-lift results to channel spend — so you know which campaigns are genuinely driving revenue, not just claiming credit. Start your free trial →
Siddharth Gangal is the founder of Fairview. He built the attribution layer after watching operators scale retargeting spend to six figures per month on attributed ROAS numbers that geo-lift tests later revealed were driven 60% by organic demand.
See it in Fairview
Track Geo-Lift Test automatically.
14-day free trial. No credit card. First data source connected in 5 minutes.