Better Data ≠ Better Conversion
Targeted ads may harm the likeliness of users to purchase the advertised product
Throughout my professional career, I have worked in roles that incorporated research and analysis. In those roles, I needed to either analyze data from predefined datasets, or decide what data to collect, collect it through open sources, and analyze it.
One of the biggest challenges was to interpret the data properly so that clients could act upon it. The interpretation process requires ignoring several data points while emphasizing others for creating a coherent line of reasoning.
In product related projects, the analysis was aimed at creating a methodology that could be later replicated for similar projects. Inductively, the approach taken for a specific use case was utilized for similar analyses.
In all of my projects, data by itself could not lead directly to conclusions, unless making assumptions and providing interpretation.
Data Accuracy is a vague term
When collecting data, one of the key principles is to define accuracy, which is usually a pre-defined and project-related term.
The philosopher Immanuel Kant has defined truth as the match between the object and its definition. For data, truth is equivalent to accuracy — as it tests whether what you wanted to collect was actually collected.
Accuracy must be defined and validated. For example, when analyzing online reviews, one should verify that the proper sources were selected and all relevant reviews were scrapped properly.
Even when data was collected properly, it can still be biased and subject to exogenous manipulations, such as paid reviews.
Even the best data requires interpretation
Interpretation is a key aspect in analyzing data. Throughout history, we have witnessed how major crises and sudden financial collapses were not predicted. In many cases, we read hindsight analyses that point out how the crisis could have been predicted by using the data at hand.
Interpretation requires applying several premises and assumptions on the situation, and avoiding other premises and assumptions for clarity. Also, predicting the future is always susceptible to mistakes.
Better data in advertising may impede conversion
In digital advertising, a lot of data is collected on users’ habits and behaviors. That data is utilized for matching the right ads to the user. The underlying assumption is that better understanding of the user would result in better match, more clicks, and eventually more revenue for both marketers and advertisers.
The utilization of data by digital advertiser collides with basic activities users are making to complete their consumer journey. Better targeting usually interferes with privacy and is more intrusive from the user’s perspective.
Specifically in purchases that include high consumer involvement, the process of research and decision making is critical. Consumers wish to validate their choices, without jumping to the first option that pops out. In that sense, even when a consumer clicks on a hyper targeted ad, it may not correlate with the purchase intent and actual purchase.
Consumer Journey includes research and interpretation
Consumers behave like analysts when purchasing online, and use their online skills to map the options, weigh them, and decide based on a variety of factors.
The first ad is not referred as the right answer, but merely as an additional option that comprises the consideration set. When information seems to be biased, consumers may doubt and ignore the first options, as these were not collected by them.
For that reason, multi-billion dollar brands have taken a step back from hyper targeting to mass reach, as the first tactic was proven to be costly and inefficient
Analysts and consumers share similarities that there is a need for accuracy, truth and interpretation before making a decision.
