Article
Platform Uncertainty and Strategic Decision-Making in Digital Advertising: An Empirical Investigation Among Meta Ads Professionals
More and more, digital advertising platforms are opaque, making it harder for advertisers to be certain of the ads they're served, the cost of those ads, and the rules that they're required to follow. Meta's advertising platform is a prime example of this trend as the platform is constantly evolving, has automated campaign tools, has some signals that are lost with privacy, and some rules for advertising that are unpredictable. This study investigates the impact of platform uncertainty on practitioners' strategic decisions for managing their Meta ad campaigns. The study builds on the theories of environmental uncertainty and resource dependence and defines platform uncertainty as algorithmic opacity, policy volatility and unreliability of measures. The 312 Meta Ads professionals surveyed were media buyers, agency strategists, in-house marketers and freelance consultants. The survey data was analyzed using structural equation modelling to test for relationships between platform uncertainty, decision-making strategies (adaptive experimentation, risk-averse conservatism and platform diversification), and perceived campaign performance. The quantitative results were augmented with semi-structured interviews with 18 respondents. The results indicate that opacity of algorithms, unreliability of measurement and volatility of policies are the factors most strongly associated with adaptive experimentation, while risk-averse behavior and channel diversification are most closely associated with volatility of policies. Those with extensive experience, and those with first-party data infrastructure, had less negative impact on perceived performance. The uncertainty-performance relationship was partly attributed to adaptive experimentation, and the degree of conservatism had a negative relationship with perceived effectiveness. Rules of thumb, peer learning networks, and "folk theories" are the ways practitioners interpret the behavior of algorithms, as revealed in interviews, but not explicitly guided by the platforms. The study expands the uncertainty framework to the case of ad dependent to the platform and proposes strategies that help shield from dependency risk. Leveraging first-party data, testing structured and diversified media portfolios are practical implications. Limitations of the current research and directions for future research are outlined, such as making a longitudinal study and a cross-platform design.