Lift Happens: Popular Tools for Accurate Campaign Impact Analysis
When your advertising efforts are up and sales are increasing, that’s great news, but it doesn’t necessarily mean the ads are the reason. To truly understand the impact of your campaigns, you need to measure lift — the incremental improvement in a key performance indicator (KPI) resulting from a marketing activity. Lift quantifies the effect your campaign made, distinguishing it from sales that would have occurred regardless. Attribution models just tell you how many people clicked on your ads, but not more than that. Some of those who clicked probably would have made a purchase anyway, while others didn’t convert despite clicking on the ads. Only lift measures the true causal impact of your campaign, comparing actual results with what would have happened without the campaign.
Measuring lift is essential because it separates the noise from the actual impact of your marketing efforts. It helps you determine which sales can be attributed to your campaign and which would have occurred regardless. Understanding lift provides a clearer picture of your campaign’s ROI and guides future marketing strategies. However, measuring lift is challenging. In reality, you can only observe one outcome — either what happened with the campaign or without it, but never both simultaneously. This makes direct calculation impossible. Thankfully, several tools can estimate lift accurately, allowing marketers to understand the true impact of their efforts.
In this article, we’ll explore three popular tools to measure the lift of a campaign. Each tool uses a different approach and can be used in various situations.
GeoLift by Meta
GeoLift by Meta is a package that performs geo tests for any platform. Geo tests involve running your marketing campaign in specific geographical areas while keeping other areas as control groups. The key is to select test areas (where you run the ads) that are similar enough to the control areas to be comparable. By comparing the performance between these groups, you can estimate the lift attributed to your campaign. Since it uses groups rather than individuals, GeoLift remains effective even with new privacy laws.
GeoLift provides causal insights by showing the direct impact of your campaign. It’s relatively inexpensive and simple to set up, making it a cost-effective option for marketers. The tool is user-friendly, with Python and R packages available for easy implementation. However, GeoLift has some limitations. It uses geographic locations as data units, which can result in a limited number of data points and larger confidence intervals. This can make the results uninformative, especially when regional variability adds noise to the data. Running geo tests can require pausing ads in some locations during the test, which can be disruptive. Additionally, geo tests provide a snapshot in time, not accounting for ad hoc changes.
For straightforward scenarios, GeoLift by Meta is the go-to tool. It assists with setup, power analysis, and test design. After performing the test, you can easily load the data and get results with a few simple lines of code. For more complex scenarios, professional assistance may be needed. For very simple scenarios, you can use GeoX by Google, a tool inside Google Ads that is very easy to use but somewhat limited.
Causal Impact by Google
Causal Impact by Google is a package that analyzes data from any platform. Causal Impact by Google uses time-series data to predict what would have happened without your marketing intervention. By analyzing trends and patterns from past data, you can create a model that estimates the expected outcome without the campaign. Comparing this prediction to your actual results gives you an estimate of the lift. This method provides a robust framework for understanding the impact of your marketing efforts over time.
Causal Impact leverages historical data to predict outcomes, offering detailed insights into your campaign’s impact. However, it requires a good understanding of time-series modeling, making it more complex to implement. The method relies on the assumption that the outcome series can be predicted by other control series that have not been impacted by the intervention. It also assumes that the relationship between the treated and control series remains consistent after the intervention. It’s essential to understand and verify these assumptions in each application to ensure accurate and meaningful conclusions. Additionally, it needs extensive historical data for accurate predictions, which can be a limitation for some marketers.
Conversion Lift by Meta
Conversion Lift by Meta is a tool that analyzes Facebook ads. It involves dividing your audience into two groups: one exposed to the marketing campaign and one not exposed (the control group). By comparing the behaviors and outcomes of these groups, you can measure the direct impact of your campaign. In the past, this was the best approach to measure lift. However, new privacy regulations make tracking users difficult, presenting a significant challenge for conversion lift studies. According to Meta’s documentation, the company addresses these issues to some extent, but we’re not sure exactly how. You provide them with the data, and they give you the results, which can feel like a “black box” process with limited control over the analysis. Additionally, it’s not a tool that can be used for other platforms, which means you can’t really compare the result of different channels.
Conclusion
Measuring lift is crucial for understanding the true impact of your marketing campaigns. The tools detailed here offer various approaches to estimate lift accurately. By understanding their strengths and limitations, marketers can make more informed decisions and maximize the effectiveness of their campaigns.
