Gartner report spotlights the four main kinds of attribution

Advertising, marketing, sales – those are only part of a marketer’s challenge. The missing piece: determining which ads or marketing were most effective in making sales or getting other desired responses. But attribution has many flavors, some of them inscrutable to many marketers.

To help sort them out, research firm Gartner has released a new report, “Use 4 Methods to Measure Marketing’s Impact” [subscribers only] that details the nature, benefits and complexities of the four main methods.

‘Less-than-perfect lens’

Analyst and report author Joseph Enever acknowledged that attribution offers “a less-than-perfect lens through which to see the real-world effects,” and added that “no single methodology provides a complete or perfect answer.”

Last-click, single touch, rule-based attribution, if focused on a single channel, doesn’t require advanced tech, the report noted, but it yields a “myopic view” of marketing impact. The picture becomes much more complicated when omnichannel marketing is involved.

Gartner recommends that marketing leaders should:

• Take an integrated approach to marketing measurement, using the most appropriate method for specific use cases, and establish realistic expectations among stakeholders.
• Make sure tools and methods are appropriate for specific purposes. In other words, brand awareness campaigns don’t benefit from conversion analysis tools.

The report covers the big four of attribution: Marketing Mix Modeling (MMM), Multitouch Attribution (MTA), Holdout Testing and Unified Measurement Approaches (UMA).

The big four of attribution

MMM is a top-down method the integrates overall sales and spending patterns to create models that determine the effect of given marketing efforts on goals, such as increased sales and revenue. Because of its complexity, MMA is often used to provide insights on a quarterly or annual basis, and its complexity requires special skills and a significant investment.

Use cases for MM include forecasting offline store sales from online investments, as well as measuring the mix of longer-term channel investments for specific goals.

MTA is bottom-up and focused on digital channels, employing user-level data like impressions and clicks to track such online marketing as paid search, display ads and video ads. It is used to identify, for instance, the most effective paid search keywords or if display ads on web sites drive sales.

Holdout Testing, also known as “test and control,” measures the incremental value of specific kinds of marketing. For example, Holdout Testing can see if removing display ads on web sites reduces sales. Use cases include proving if a new marketing channel investment is worthwhile, such as addressable TV.

Finally, Unified Measurement Approaches (UMA) combines the top-down MMM and the bottom-up MTA. It is intended to integrate MMA’s person-level model with MMM’s large-scale approach.

The report recommends that marketers use the most appropriate attribution method for specific goals and channels, combine them as necessary, and understand the limitations and resource needs of each.