Why Geo-Lift Testing Alone Falls Short—And How Stella Ensures Accurate Incrementality Testing

In recent years, marketers have been increasingly focused on measuring the true impact of their campaigns in the face of cookie-based tracking limitations and evolving privacy regulations. While many look to geo-lift testing and incrementality testing as potential solutions, these methods, if done incorrectly, can produce flawed results. At Stella, we recognize these challenges and offer a more reliable, comprehensive approach to incrementality testing that helps marketers gain actionable insights without falling into common pitfalls.

The Problem with Relying on Geo-Lift Testing Alone

Geo-lift testing, often promoted as a next-generation solution for measuring incremental ad impact, relies on creating a "synthetic control" region to simulate what would have happened if ads had continued in a given test area. While synthetic controls aim to mimic real-world conditions, they can introduce a significant amount of bias if not carefully managed. The idea of blending data from different regions using weighted averages sounds appealing but comes with inherent risks. If done improperly, the synthetic control group may not accurately reflect what would have happened, leading to skewed results.

Why Stella Is Different: Personalized, Hands-On Support

One of the primary concerns with any form of incrementality testing is whether the results can be invalidated due to errors in experimental design or execution. This is where Stella stands apart. Our business model is built around ensuring that tests are done correctly, every single time. We work 1-on-1 with your team to ensure that the right regions, channels, and control groups are selected, and any bias is minimized.

Stella doesn't just provide a platform—we provide expert guidance to ensure that the testing methodology is sound, meaning your results are reliable. Whether you're testing incrementality at the channel, campaign, or even tactic level, we help set up tests that deliver statistically significant results, allowing you to make confident, data-driven decisions.

Statistical Significance and Why It Matters

A key factor in incrementality testing is reaching statistical significance—proving that the observed results are due to the marketing intervention and not random chance. Achieving this requires a sufficient sample size, proper randomization, and consistent test design. At Stella, we recommend that businesses spend a minimum of $85k per month per channel or achieve 100+ orders per day in a single region before venturing into incrementality testing. This ensures that the data collected is robust enough to reach statistical significance, making the test results reliable and actionable.

Stella's proprietary models analyze your historical sales data from the past 90 days to identify the best regions to hold out, ensuring that the test achieves significance with the lowest possible investment and the shortest amount of time.

Geo-Lift Testing: A Useful Tool—But Not in Isolation

While geo-lift testing has its merits, it's essential to understand that it’s not a perfect or foolproof solution. Stella recognizes that geo-lift testing alone isn’t enough to provide a complete picture of campaign performance. We recommend using geo-lift testing as one of many tools in a broader multi-faceted approach that includes:

• Multi-Touch Attribution (MTA): To track customer journeys and measure touchpoints.
• Self-Attribution: Through intake forms to get direct insights from your audience.
• Media Mix Modeling (MMM): To optimize your media spend across all channels.
• Incrementality Testing: To measure the true lift your campaigns provide.

By incorporating a range of measurement tools, Stella helps ensure a well-rounded understanding of campaign effectiveness. No single method—whether it’s geo-lift testing, MTA, or MMM—can give you the full story. But by combining these approaches, we provide marketers with a comprehensive measurement framework that drives growth.

Common Pitfalls—and How Stella Avoids Them

There are several common mistakes businesses make when running geo-lift tests, particularly when trying to conduct them in-house or relying on platforms that don’t provide the right level of expertise. These include:

1. Poor Test Region Selection: One of the biggest errors is selecting test and control regions without thoroughly evaluating how similar they are in consumer behavior. At Stella, we focus on finding regions with high correlations in purchasing behavior, ensuring accurate test results.
2. Seasonal Bias: For seasonal businesses, running tests during high-demand periods (like Black Friday) can throw off the results. Stella accounts for seasonality in our models, adjusting synthetic controls to avoid misleading conclusions during extreme sales fluctuations.
3. Incorrect Budget Allocation: Another major issue is incorrect budget allocation during the test period. Excluding test regions but not adjusting the budget can result in inflated control region ad spend, distorting the test. Stella ensures that budget distribution is handled properly, so you get a clean, unbiased view of your campaign’s incremental impact.
4. Bid Strategies and Synthetic Controls: While using Target CPA or Target ROAS bid strategies can complicate testing, Stella adjusts its model fit based on historical performance data for each region. We account for variations in these bidding strategies so you can still run accurate tests without skewing your synthetic control group.

Comprehensive Testing Beyond Geo-Lift

Most of Stella’s clients are testing incrementality at the channel and campaign levels, but we go deeper. Our tool allows you to test specific tactics and bid strategies, such as whether “Max Conversion Value” is more incremental than “Target ROAS” on Google or whether “7-day click” attribution is more effective than “7-day click, 1-day view” on Meta retargeting campaigns.

Stella is designed specifically for mid-market ecommerce brands, helping you test incrementality in as little as 15 days while ensuring statistical significance. Our approach helps advertisers understand which channels, campaigns, and tactics are driving true incremental value, allowing you to optimize budgets without increasing spend unnecessarily.

The Bigger Picture: A Holistic Approach

At Stella, we don’t advocate for relying solely on incrementality testing. We understand the importance of using a blend of measurement tools—Incrementality Testing, Media Mix Modeling, Multi-Touch Attribution, and Self-Attribution—to get a complete picture of your marketing performance. Each approach has its strengths and weaknesses, but together, they provide the insights you need to make data-driven decisions and deliver real value for your clients.

In a world where privacy regulations and tracking limitations continue to evolve, it’s more important than ever to have a reliable and flexible measurement strategy. Stella offers a unique combination of incrementality testing and media mix modeling, giving you the tools you need to optimize performance while keeping costs in check.

Conclusion: Stella Delivers Incrementality Testing Done Right

Incrementality testing, when done correctly, can provide incredibly valuable insights into the effectiveness of your marketing efforts. The key is doing it right—minimizing bias, ensuring statistical significance, and using it in conjunction with other measurement tools. Stella offers more than just a platform; we provide expert, personalized guidance to help you set up and execute tests that yield accurate, actionable results. With our approach, you can confidently show your clients the channels, campaigns, and tactics that are driving growth, and look like a rockstar while doing it.

Let Stella be your partner in incrementality testing and media mix modeling, and start making the kind of marketing decisions that will drive long-term success.

‍