With stricter regulation over consumer privacy and data collection, it has become increasingly difficult to track the way in which consumers interact with brands across a multifaceted, and sometimes fractured, marketing landscape. Robust analytical and modeling solutions are being leveraged to help brands understand the impact of their marketing efforts on raising awareness and driving profitability. At the core of these solutions is attribution measurement. In this article, we will define attribution as a collection of analytical techniques used for assigning credit to products, channels, campaigns, countries, markets, or any other relevant marketing stimulus for driving consumer actions/behaviors. The goal of attribution measurement is to identify the impact associated with these marketing stimuli to better understand how the components of a marketing strategy are working and to use these findings to optimize toward the most efficient results per marketing dollar.

What are the limitations of current attribution measurement methodologies?

Currently, marketers utilize disparate measurement methodologies to understand marketing performance and assign attribution. However, ​no single methodology can answer all the questions associated with attribution, and they can sometimes provide opposing answers when analyzed in silos. As George Box, a British statistician, famously said: “All models are wrong, but some are useful.” So let’s better understand the various methodologies that have been developed for attribution measurement:

  • Marketing mix modeling (MMM) quantifies the ROI impact of channel-level performance by analyzing historical marketing-specific factors and external factors that lead to varying results based on media mix. The limitations to this top-down approach are that it looks at aggregate-level macro effects but does not account for user-level interactions and is not as suitable for measuring addressable marketing.
  • Full funnel modeling (FFM) expands upon MMM measurement to understand the impact of media (and other drivers) at each stage of the consumer purchase path, as well as the way in which each part of the sales funnel affects the others. Limitations are similar to that of MMM.
  • Multi-touch attribution (MTA) uncovers the relationship between fragmented channels by analyzing trackable digital data across marketing touchpoints that produce relevant consumer interactions. The limitations to this bottom-up approach are that it relies on digital data, which have low coverage and are becoming increasingly difficult to capture at the individual level, and it is less suitable for measuring non-addressable marketing, offline media, and external factors.
  • Advanced A/B testing and closed-loop experiments evaluate the incremental lift in campaign performance as a result of sequential changes to the marketing strategy. The key limitation to this approach is that it does not consider macro effects, particularly on ROI.

To overcome the limitations of any singular measurement methodology, marketers must implement a unified solution that combines these approaches to produce a holistic view of marketing performance at both macro and granular levels in order to drive the most informed optimization decisions for business planning.

How does a unified measurement approach come together?

The key benefit of a unified measurement approach is the ability to answer an organization’s most crucial and complex questions that impact all stages of business planning, from in-flight budget allocations to annual budget planning to quarterly updates and re-allocations across lines of business, countries, campaigns, and channels. A unified attribution measurement approach puts consumer impact at the center of the optimization strategy, regardless of where in the organizational matrix the budget sits; it aims to provide a holistic strategy to drive all-up performance for the organization.

How does a marketing organization go about implementing a unified attribution measurement solution? These are some key considerations and recommendations for implementing a unified approach:

  • Establish a team of analysts and marketers with expertise in attribution, planning, media buying, and execution who work together to evaluate performance and implement informed optimization strategies; ensure performance measurement is used to drive actionable improvements.
  • Identify and communicate the relevant KPIs and metrics that determine successful performance for your organization, be it brand awareness, profitability, or a combination thereof.
  • Develop a unified data collection, data sharing, and modeling strategy that enables the discovery of blended results across various data sources and measurement methodologies, thereby mitigating the limitations of disparate approaches while leveraging the strengths of each.
  • Utilize both top-down and bottom-up attribution measurement tools that provide actionable insights for both granular execution optimizations (for example, search term efficiency or path to conversion analyses) as well as an understanding of ROI impact due to macro effects (both internal and external) across lines of business, channels, countries, markets, etc.
  • Measure not only the gross impact of marketing but also the incremental impact by employing robust testing strategies with statistically measurable results, iterate upon these experiments using historical results, and combine findings from testing with attribution modeling.
  • Take a consumer-centric view by incorporating insights that analyze consumer behavior at all stages of the purchase path and understand how every part of the sales funnel is inter-related. It can also be beneficial to go one step further and deploy a market research study to survey consumers directly.
  • Align annual/quarterly/in-flight budget strategies with ongoing performance measurement and scenario planning coming out of joint attribution tools and break down budgetary silos to drive overall profitability for the organization.
  • Invest in MarTech and MarSci platforms, such as Kinesso’s Outcome Planner, that provide unified measurement tools to inform budget optimization at the marketing levels relevant to your organization (i.e. by LOB, channel, market, etc.).

What questions can marketers answer using a unified measurement approach?

Let’s examine a use case to illustrate the implementation of a unified measurement approach to answer key attribution questions related to marketing impact on sales. First, determine the performance goal; in this case let’s assume the goal is to maximize profits through efficient customer acquisitions across online and offline marketing channels, thereby resulting in ROI growth. Next, we identify the relevant attribution tools to leverage for determining channel-level performance as well as understanding consumer behaviors. Using a combination of MMM/FFM and MTA models is a good foundational layer, but they must not be applied in isolation of one another. Therefore, we could utilize a machine learning blended model that merges: 1) channel-level ROI quantification, media mix results, and sales funnel optimizations generated by MMM/FFM, and 2) conversion path predictions generated by MTA using individual-level log data. Each attribution model is given a situational weight depending on its predictive strength to offset weaknesses in any singular methodology and fix model biases, which subsequently results in more accurate and balanced predictions that lead to robust, well-rounded optimizations. We can then layer in additional A/B and closed-loop experiments to test various treatments against a given audience and measure the incremental impact of changes made to the marketing strategy. Experiment results are compiled for meta-analysis and optimization in conjunction with attribution modeling. Using this holistic approach, we can then answer questions such as:

  • What is the best way to allocate a $100MM budget across channels to maximize sales?
  • How do we allocate this budget across countries/markets/lines of business?
  • What is the optimal media mix if I can only spend $5MM on addressable TV?
  • How do we minimize losses if the budget is reduced by 20%?
  • Which flighting strategy should we employ to remain most efficient for the year?
  • What is the projected sales impact for the second half of the year based on the proposed media strategy?
  • What is the effect of our branding efforts on consumer awareness and consideration? And how does it impact our ROI?
  • Is there a preferred digital conversion path for a given audience, and how do we allocate our digital media budget across channels to maximize sales?
  • Can we simulate budget requirements based on an ROI goal of 150%?
  • Which marketing treatment yielded the most significant incremental lift in conversions, and how did this affect our overall ROI?

Overall benefits of unified attribution measurement

When multiple attribution methodologies are combined into a unified measurement approach, marketers can produce performance-driven optimizations that account for impact on ROI by strategic groupings (i.e. channel, line of business, etc.), consumer behaviors across the entire sales funnel, and their impact on profitability, with full coverage across all media activity and always-on tracking for ongoing optimization. With these tools consolidated into a single solution, organizations have a more holistic understanding of marketing impact that facilitates their ability to plan and allocate budgets across channels, lines of business, audiences, campaigns, creatives, and other marketing touchpoints. Enabling marketers to leverage these tools for media planning and execution using more informed budget allocations will subsequently lead to increased efficiency and generate profitable growth for the organization, while keeping a consumer-centric view.