As signal loss and regulatory changes make media more difficult to measure, changing the marketing mix is critical, according to Gartner's Matt Wakeman. But the strategy is more effective if some specific optimization tactics are used.
When my wife and I were paragliding in Lima, our guide was also a business owner. After the flight, we started talking about our trip and how we chose your company. We told him that we had heard good things about our hotel (obviously sellers hear this by word of mouth).
I couldn't pass up the opportunity to continue practicing my young Spanish, so I asked him what he thought about travel guides and apps that allow local businesses to advertise for travelers. He shrugged and said he didn't use them, mainly because he didn't know if the publicity was worth it.
It's certainly difficult to measure such an impact for a local family business, let alone a word-of-mouth impact. However, even global brands with billions of dollars in sales make marketing difficult to measure.
While measurement is critical to demonstrating marketing value to your organization and optimizing campaigns, channels and tactics, this has become increasingly difficult in recent years due to a number of key factors.
First, efforts are more indirect: marketing influences customer attitudes, which then trickle down to customer actions and purchase outcomes. And an unstable and uncertain world only makes the effects even more indirect.
Second, and most importantly, out-of-date data, third-party cookie opt-outs, and regulatory changes hinder marketing measurement. These realities negate approaches directly related to digital advertising and measurement, such as multi-touch and post-view attribution (Google offers over 6.9 million results due to poor data alone, so we won't discuss that here).
Marketers and measurement vendors have turned to the marketing mix model (MMM) as a solution to these problems. MMM measures the impact of advertising, promotion and branding across all channels, taking into account the impact of factors external to the brand, such as competitive activity or consumer sentiment. When used properly, it is an important tool for budget decision-making, and they are becoming more accurate as the threat of economic recession approaches.
Examples of mixed model analysis include: predicted impact of online marketing spend on brick-and-mortar sales; optimal delivery patterns or sequences and the amount of advertising exposure required to increase sales; and the optimal overall combination of online and offline advertising costs for a given target.
MMM is protected from factors such as data aging because it uses aggregated time-series data as input, such as the number of views per day for a given campaign in a given geographic area. Methods based on digital path assembly have more and more gaps.
Major brands are recognizing MMM as an increasingly important measurement tool as marketers need to assess the impact of new channels, the evolving digital tracking environment and changing consumer behavior. The most effective marketers combine MMM with incremental methods.
In marketing, incremental measurement quantifies the unique impact a campaign, experience, tactic, or ad had on one or more business outcomes. Measuring incrementality can be done using a variety of methods, from relatively simple obscurity tests to more complex universal control groups and advanced marketing mix models. All three methods have the same goal: to isolate changes caused by marketing.
For example, advertising typically compares the behavior of consumers who are shown ads and those who are not. The difference is a measure of the increase in advertising. The analyzes they perform in this way ensure that the two groups are as similar as possible. The only thing that should differ between the two is the marketing exposure.
How do the best brands do it? In general, they have three main approaches:
1. Check the reliability of MMM forecasts
Treat the marketing mix model as a strong proposition, not as a prescription to be followed. How should marketers treat these offers, if not as prescriptions? Run tests to verify the model's predictions.
An example of this is co-marketing testing, where marketers change spend in one market but not the other, then wait and compare how the results change between the two markets. If the test market increases sales or profits more than the control market, this result supports the MMM recommendation. In this case, the model has passed the test and its impact on marketing decisions should increase.
2. Use Modeling and Optimization to Create Better Marketing Plans
In fact, MMM is a forecasting tool. Most comprehensive marketing companies offer an interface or simulation tool. In addition to being able to adjust the marketing levers, the scenarios can change the external factors included in the model. These planning tools attempt to account for marketing elasticity and the fact that each marketing action is on a different part of the response curve.
3. Treat the dimensions of strategic marketing as a journey
The measure will continue to change as web browsers, walled gardens and regulators develop their policies. As brands become analytically mature, they often need different information. Both of these forces should eventually lead to changes in the marketing dimensions of organizations.
Marketers should focus on long-term improvements, such as testing new data sources or deep analytics to better understand specific channels.
For large advertisers, aging data hampers digital marketing measurement efforts. While traders and the measurement industry have returned to MMM, traders who take MMM at face value lose much of their value if they fail to follow these steps.
Matt Wakeman is a senior marketing analyst at Gartner.
