4/30/2023 0 Comments Web advertising postview![]() Our design provides faster and cheaper results including double-blind, to the users and to the serving engine, post-auction experiment execution without ad targeting bias. We present a novel randomized design solution for incremen-tality testing based on ghost bidding with improved measurement precision. Similarly, ITT and ghost bidding solutions provide greatly decreased precision since many experiment users never see ads. A fundamental challenge with these is that the serving engine as treatment administrator is not blind to the user treatment assignment. Current literature and industry practices to run incrementality experiments focus on running placebo, intention-to-treat (ITT), or ghost bidding based experiments. Running randomized controlled experiments is the gold standard in marketing incrementality measurement. Measuring the incremental value of advertising (incrementality) is critical for financial planning and budget allocation by advertisers. Whether or not you share our perspective completely, we hope we facilitate future research in this area by pointing to related articles from multiple contributing fields. Our perspective is that these observations open up substantial fertile ground for future research. This observation helps to explain at least one broad common CDM practice that seems "wrong" at first blush: the widespread use of non-causal models for targeting interventions. Finally, (3) causal statistical modeling may not be necessary at all to support CDM, because there may be (and perhaps often is) a proxy target for statistical modeling that can do as well or better. The upshot here is that for supporting CDM it may be just as good to learn with confounded data as with unconfounded data. (2) Confounding does not have the same effect on CDM as it does on CEE. (1) We should carefully consider the objective function of the causal machine learning, and if possible, we should optimize for accurate "treatment assignment" rather than for accurate effect-size estimation. We draw on recent research to highlight three of these implications. Technically, the estimand of interest is different, and this has important implications both for modeling and for the use of statistical models for CDM. Our experience is that this is not well understood by practitioners nor by most researchers. This article highlights an important perspective: CDM is not the same as CEE, and counterintuitively, accurate CEE is not necessary for accurate CDM. Recently, we have seen an acceleration of research related to CDM and to causal effect estimation (CEE) using machine learned models. For example, businesses often target offers, incentives, and recommendations with the goal of affecting consumer behavior. The practical bottom line: evaluating campaigns and optimizing based on clicks seems wrongheaded however, there is an easy and attractive alternative|use a well-chosen site visit proxy instead.Ĭausal decision making (CDM) at scale has become a routine part of business, and increasingly CDM is based on machine learning algorithms. Specifically, predictive models built based on brand site visits do a remarkably good job of predicting which browsers will purchase. ![]() The good news is that an alternative sort of proxy performs remarkably well: observed visits to the brand's website. The bad news is that across a large number of campaigns, clicks are not good proxies for evaluation nor for optimization: buyers do not resemble clickers. The most commonly cited and used proxy for success is a click on an advertisement. The paper presents bad news and good news. Proxies are necessary because data on the actual goals of advertising (e.g., purchasing, increased brand affinity, etc.) often are scarce, missing, or fundamentally difficult or impossible to observe. Measuring success is critical both for evaluating and comparing different targeting strategies, and for designing and optimizing the strategies in the first place (for example, via predictive modeling). This paper presents results across dozens of experiments within individual online display advertising campaigns, each comparing different 'proxies' for measuring success. However, there often are too few purchase conversions for campaign evaluation and optimization, due to low conversion rates, cold start periods, and long purchase cycles (e.g., with brand advertising). A main goal of online display advertising is to drive purchases (etc.) following ad engagement.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |