How Focusing on True Value Empowers Modern Marketing Measurement

As members of the Ads Research team at Facebook, we’ve been studying incrementality as part of our mission to understand the impact of new concepts and techniques in the advertising industry. Recommended by leading marketers and validated through researchers at Kellogg School of Management at Northwestern, focusing on incrementality measurement enables you to make better data-driven marketing decisions, while still being able to access precise insights in privacy-friendly ways.

What exactly is incrementality and why is it so powerful? And how can you craft an incrementality-focused measurement strategy?

Here's what you need to know.

Incrementality is the additional business value generated as a direct result of a marketing campaign or media exposure. Typically, marketers struggle to determine if their campaigns actually drove impact, like sales or brand awareness: Would a sale have taken place regardless of whether someone saw your ad? As the ad ecosystem and people’s behavior continues to evolve, it’s more important than ever to understand marketing performance and make informed marketing decisions.

Incrementality measurement helps you determine how and where your marketing efforts are contributing to bottom-line success. It uses experimental methods to compare the behaviors of people who see or don’t see your ads, allowing you to identify the impact on outcomes such as conversions. Effectively adopting it requires not just picking specific solutions but also building a broader marketing culture dedicated to testing and learning.

As marketing evolves and the ways advertisers can use data changes, incremental measurement lets advertisers test their current strategies and ensure they’re still effective and evaluate the impact of new techniques and tools. At the same time, measurement is evolving, and how marketers measure incremental value may need to change. As measurement tools and solutions are impacted by and adapt to industry changes around data availability, marketers should consider multiple solutions and make sure to match the right solution to their goal.

According to our research,¹ incremental and nonincremental measurement approaches disagree on a winning marketing tactic nearly 25% of the time. And when marketers choose the wrong tactic, businesses lose out on an average cost-per-action (CPA) improvement of 64%.¹

In other words, understanding incrementality is important because it connects marketing activity to business value, thereby identifying what drives results most efficiently. Armed with these data-driven insights, you can make the strategic marketing decisions that will have the greatest measurable impact on your bottom line.

Experiment-based solutions are the gold standard for measuring incremental business value, but they might not be practical in every scenario. Sometimes you may have constraints such as budget or time limitations that necessitate using less precise approaches.

If you can’t run an experiment, you can still ensure your measurement methods are as close to measuring incrementality as possible. You can think about both metrics and methods along a spectrum of incrementality, spanning from nonincremental approaches like rule-based attribution all the way to experiment-driven solutions like randomized controlled trials (RCTs). The closer you can get to experiments, the better off you are, but there are many valuable methods available along the way.

Also, keep in mind that using a variety of solutions has benefits; a multipronged approach can help to reduce confirmation bias and validate the accuracy of your measurement strategy.

Just as measurement solutions can be evaluated along a spectrum of incrementality, so too can outcomes and metrics.

To inform strategic decisions, it’s best to measure the true business outcomes you are trying to drive, like sales, as they are more directly tied to the bottom line compared to metrics such as clicks.

However, as with experiments, this may not always be possible. If you can’t measure the business outcome itself, the alternative is to see if another metric, such as clicks, can serve as a proxy. You can validate this by testing your proxy metric to see how it ties to the ultimate desired goal. Again, the aim is to get as close as possible to anchoring your marketing measurement in incremental business value.

Once you’ve assessed the measurement solutions and outcomes, you can put it all together and see where you land along the spectrum of incrementality.

For example, take marketing mix models (MMMs), which use statistics to estimate how your marketing is impacting sales. Calibrating an MMM with experiments provides a relatively high level of incrementality, moving it to the more incremental side of the spectrum, whereas not calibrating an MMM with experiments moves it to the less incremental side. The more incremental an MMM is, the closer it will come to measuring true business value of your marketing activities—and the more confident you can feel in using it to make key decisions.

Crafting a measurement strategy founded on incrementality is not a one-off endeavor, but an ongoing, iterative process. The key steps along the way include:

• Identify your business goals: Start with understanding your desired business outcomes and which business decisions you need to make;
• Consider constraints: Identify limitations that may prevent you from using the most incrementality-based solutions and metrics;
• Select outcomes and solutions: Based on your goals and constraints, determine which metrics and business outcomes and approaches best fit your needs along the spectrum of incrementality; and
• Assess and adjust: Observe, test and learn to ensure your measurement is accurate and that you are accessing the insights you need to make strategic decisions.

Ultimately, the goal should be to continually reexamine and refine your strategy so that it is anchored as much as possible in true business value. By doing so, you’ll ensure your measurement is strategically valuable and future-proofed.