With the growth in retail media spending, questions are now being raised about how best to measure the effectiveness of retail media with the limited available data in a new era of privacy. While last touch attribution is widely used today, it is not the only available methodology for retail media measurement. There are multiple alternatives that marketers can consider to effectively measure campaigns. While these approaches present the potential for improvement, each of these has upsides and downsides.
Here’s a look at three different measurement tools:
Direct Attribution
Marketers have sought to build on last-touch attribution to create more nuanced ways of apportioning credit to an ad for a sale. These include methodologies such as multi-touch attribution, linear attribution, and algorithmic attribution.
With the rise of retail media, new approaches have emerged to apply these methods, such as customer data platforms and clean rooms. These offer granular customer, event, and order level data that can be used to reconstruct a user’s path to purchase and assign credit to each interaction using approaches other than last touch.
The strengths: These tools assign credit across the measured customer journey, and can be very predictive.
The limitations: These tools still offer an incomplete picture. There are still gaps in the ability to build a complete picture of data due to the challenges presented by walled gardens and consumer privacy detailed above.
Media Mix Modeling (MMM) or econometrics
Also known as econometrics, this approach involves creating multi-factor regression models that are used to estimate the impact each touchpoint or factor has on sales. These can include external factors like weather and seasonality, as well as marketing channels outside of retail media.
The strengths: By providing the ability to integrate marketing channels outside of retail media, MMM can provide a comprehensive picture of not only the customer journey but all of the factors that lead to a sale. MMM is particularly good at simulating “what if” scenarios, such as how budget increases will impact ROAS over time.
The limitations: MMM tends to be slow and resource intensive, making it difficult to make decisions at the pace of the internet.
Lift analysis or incrementality A/B testing
These tests quantify the incremental return and causal effect of a particular marketing channel or campaign on an independent variable, usually sales. They are typically performed as ad-hoc tests, which provide performance reporting after the marketing activity is complete.
The strengths: These tests are designed to establish and quantify the causal connection through testing between a dependent and independent variable.
The limitations: Not every platform will provide the ability to separate clean tests and control groups. These also tend to be run in an ad-hoc nature, as opposed to a continuous and sustained campaign that meets the demands of the fast-moving digital environment
Key factors to consider when exploring alternative approaches to retail media measurement
Despite fast growth, this form of advertising remains nascent and can benefit from evolution. Alternative approaches to attribution can help to solve the issues detailed above.
When considering these alternative approaches, leaders should keep these factors in mind:
The business questions that are most important to you.
What decisions are you using the tool to support? Will it only be used to support campaign optimization, or are broader questions of overall budgeting and planning important?
The ability to integrate multichannel marketing.
Does your brand have significant marketing investments outside of retail media? Do any of these drive traffic toward digital storefronts? If so, measuring or at least controlling for these external factors may be important.
The ability to control for external factors.
Is your category seasonal or sensitive to macroeconomics? Do you have substantial brand equity already that might attract users organically to your products? If so, controlling for these external factors may be important.
The frequency of measurement.
At what frequency do you need reporting for the insights to be actionable? Do long lag periods between action and measurement erode the value of the measurement? If so, you may need an approach that can be implemented more frequently.
The available data science resources.
Do you have internal data science resources that can be dedicated to developing or maintaining a measurement solution? If you look to partner with an external party, do you have the internal resources needed to be able to support the engagement?
Is your data ready?
Access to cleansed and normalized data will be required in all of these approaches. Do you have automated data collection, cleansing, and normalization in place? Have you already integrated retail and marketing data into your data lake?
All of these approaches as well as answers to the above questions about the needs of your organization will play a part in determining the approach that is appropriate for your brand. If you’d like to talk with an expert about your brand’s needs and how to navigate retail media measurement, please get in touch.
