Marketing Measurement Overview

Effective marketing measurement is crucial for understanding the true impact of your campaigns and optimizing your marketing spend. This guide will help you understand how to measure marketing incrementality and make data-driven decisions.

Common Marketing Measurement Questions

Understanding Incrementality

What is Incrementality?

Incremental impact of marketing is the amount of conversions (e.g., signups, leads, orders, app installs, etc.) that would not occur without a the marketing. In other words, it's the amount of conversions that you would have lost without the marketing.

Why Measure Incrementality?

Incrementality provides a factual rather than subjective view of marketing performance. It is by definition the kind of measurement or attribution that marketers should care about.

However, due to simplicity and easy-of-implementation, attribution methods such as first-touch, last-touch, and cookie tracking-based attribution are more widely used.

Why Traditional Attribution Methods Are Unreliable

Increasingly restrictive privacy regulations have made tracking-based attribution methods broken:

• Reliance on Privacy-Sensitive Tracking: These methods depend on individual-level tracking (e.g., cookies, device IDs) to function. However, increasing privacy measures—such as tracking prevention in browsers (ITP, ETP), app data restrictions (ATT on iOS), and user consent requirements—limit or block this data, making tracking and the resulting attribution incomplete.
• Bias Toward Conversion-Proximate Channels: These methods favor marketing channels or tactics that are easier to track, such as brand search, which happens close to conversion. As a result, they greatly underestimate the impact of upper-funnel channels (e.g., video ads, social media, connected TV, etc.) that drive demand (such as traffics to search) but lack click-through conversions.

Inrementality Measurement Approaches

Geo-based Hold-out Experiment (a.k.a Incrementality Test)

Geo-based holdout experiments, often referred to as "incrementality test" or "geolift", are an intuitive and simple way to measure incrementality but have limitations in scalability.

How it works: A geo-based holdout experiment measures incrementality by establishing a holdout group.

• Assign geographical areas (e.g., regions, states, cities, or DMAs in the United States) to test and countrol (hold-out) groups, ensuring they are similar before the test.
• Launch the marketing activity to be measured (e.g., a channel, a tactic, or a campaign) only in the test group.
• After the test is completed, compare conversion(s) between the test and control groups to estimate the incremental impact of the marketing activity.

Benefits:

• Privacy-friendly: Does not need individual-level tracking.
• Incremental: Measures incrementality by establishing a counterfactual.

Note: User-level holdout experiments, often provided as "Conversion Lift Testing" by some advertising platforms, still rely on cookie tracking and face the same limitations as traditional attribution methods described above.

Limitations:

• Can only test one, or at most, a couple of channels, tactics, orcampaigns at a time
• Requires turning off advertising in the hold-out group
• A point in time measurement which does not provide ongoing insights

MMM (Marketing Mix Modeling)

MMM provides a scalable, privacy-friendly approach to incrementality measurement. Depending on the specific approach or vendor you choose, implementation complexity and the depth of insights can vary.

How it works: MMM measures incrementality by modeling fluctuations in marketing activities in relation to conversions using statistical and machine learning models. It leverages historical data without the need to set up holdout experiments. The specifics can vary significantly depending on the vendor you choose. See the Maxma approach in the Maxma MMM section for more details.

Benefits:

• Privacy-friendly: Does not require individual-level tracking. Requires only aggregated time series data.
• Scalable: Measures the entire marketing portfolio simultaneously without requiring experiment setup.
• Incremental: Captures incrementality by analyzing fluctuations.

Limitations:

• Implementation: Requires sophisticated statistical and machine learning models and collection of data across all of your marketing channels. Depending on the specific solution you choose, it may require a significant amount of time and effort to set up. But we've made it a plug-and-play solution through Maxma MMM.
• Frequency and granularity of insights: Traditional MMM often lacks granularity and is typically conducted infrequently—usually once a year—for high-level insights. However, our Maxma MMM uses proprietary algorithm which enabls ongoing tactic level incrementality measurement.