Big Tech companies should understand the difference between prediction and causal inference and its significance for resource allotment.
Many Big Tech companies have developed platforms that provide businesses with tools and services to help them target online customers. While they can be beneficial, organisations that use them must exercise caution, as the claims are often oversold to entice brands to spend more money on ads.
While Big Tech companies like Facebook, Google and Twitter do set a precedence on online advertising, their training materials and tools available for use for other (perhaps) small businesses are inherently designed keeping Big Tech’s interests in mind. For example, Facebook's A/B testing service isn't textbook A/B testing, because not all from the groups will be tested, and those that will be aren’t treated to the same ad material.
The platform’s machine-learning algorithm only serves ads based on what it learns about a demographic’s preferences, thereby invalidating the A/B test itself. Small and medium businesses using Big Tech’s ad platforms and tools must be wary of “apples-to-oranges comparisons” that support “causal inferences.
[6 minute read]
