The Role of Programmatic Media
Posted: Mon Jan 20, 2025 5:28 am
It is worth highlighting that the so-called " programmatic media " only works in campaigns with large investments, due to its nature of testing/validating so that artificial intelligence ( machine learning ) directs the campaign in the best direction.
In a scenario where there are many hypotheses to be validated at the same time (channels, keywords, times, regions, etc.), only a sufficiently large sample reduces the risk of wrong decisions which, in the case of programmatic media, means reducing the risk of investing incorrectly and being left without ROI .
Therefore, doing programmatic media with a small event planner email list campaign/sample, that is, less than R$100,000 per month , means putting the ROI at the mercy of chance in the sampling.
To better illustrate what I'm talking about, I'll share the example that our COO here at Hubify, Oscar Sigaki, uses:
“I have R$10,000 and would like to do a test using programmatic media (remembering that programmatic media does not replace Paid Search, it is complementary). To do this, it is necessary to do several A/B tests with dimensions such as:
Ad Exchange;
DMP / Data Provider;
Demographic, interest, and category segmentation;
Websites;
Viewability;
Creative;
between others.
With this, we have infinite combinations of the dimensions mentioned above to test, which becomes unfeasible with low investments.”
Always be aware
Here, if we need people to deal with numbers, it is important to pay extra attention , since even the best professionals in exact sciences are constant victims of heuristics (when our brain replaces a difficult question with a quick answer) and biases (systematic errors that repeat themselves in a predictable way in particular circumstances), making wrong decisions. So, imagine those who are not familiar with exact sciences.
The Law of Small Numbers
If you still don't agree with me, I invite you to reflect on this excerpt from the book "Thinking, Fast and Slow" , by Daniel Kahneman, Nobel Prize winner in Economics, in which he discusses the Law of Small Numbers and "statistical intuition".
In a scenario where there are many hypotheses to be validated at the same time (channels, keywords, times, regions, etc.), only a sufficiently large sample reduces the risk of wrong decisions which, in the case of programmatic media, means reducing the risk of investing incorrectly and being left without ROI .
Therefore, doing programmatic media with a small event planner email list campaign/sample, that is, less than R$100,000 per month , means putting the ROI at the mercy of chance in the sampling.
To better illustrate what I'm talking about, I'll share the example that our COO here at Hubify, Oscar Sigaki, uses:
“I have R$10,000 and would like to do a test using programmatic media (remembering that programmatic media does not replace Paid Search, it is complementary). To do this, it is necessary to do several A/B tests with dimensions such as:
Ad Exchange;
DMP / Data Provider;
Demographic, interest, and category segmentation;
Websites;
Viewability;
Creative;
between others.
With this, we have infinite combinations of the dimensions mentioned above to test, which becomes unfeasible with low investments.”
Always be aware
Here, if we need people to deal with numbers, it is important to pay extra attention , since even the best professionals in exact sciences are constant victims of heuristics (when our brain replaces a difficult question with a quick answer) and biases (systematic errors that repeat themselves in a predictable way in particular circumstances), making wrong decisions. So, imagine those who are not familiar with exact sciences.
The Law of Small Numbers
If you still don't agree with me, I invite you to reflect on this excerpt from the book "Thinking, Fast and Slow" , by Daniel Kahneman, Nobel Prize winner in Economics, in which he discusses the Law of Small Numbers and "statistical intuition".