Algorithms in Advertising: Should You Tell Clients About AI?
Algorithms in Advertising: Should You Tell Clients About AI?
The use of algorithms in advertising raises a natural question for marketers: should they openly declare the use of artificial intelligence in communications? A study published in the journal Nature Human Behaviour showed: personalized risk messages significantly increase conversion, while mentioning AI does not affect user decisions at all. This conclusion is critically important for specialists who fear «scaring away» the audience with technologies in advertising creatives.
What the Researchers Discovered
A team of scientists conducted a large-scale experiment consisting of three randomized field trials. More than 90,000 patients of the US healthcare system participated. The goal was to find out how messages about personal risk of flu complications stimulate people to get vaccinated.
Key Effectiveness Metrics
| Group | Result (vaccination rate) |
|---|---|
| Control (no messages) | Up to 14.7% |
| Simple reminder | Baseline level |
| High risk notification | Increase of 3.3–5.4% |
- Role of AI: Mentioning that the forecast was made by an algorithm, or providing explanations for it («explainable AI»), did not yield statistically significant differences in effectiveness.
Detailed study protocols are available on ClinicalTrials.gov.
How Much to Trust This
This is a classic randomized controlled trial (RCT) — the «gold standard» for evaluating effectiveness. A sample of 90,000 people provides high statistical power. It is important to note: this is a real field trial, not a laboratory survey. The results show real human behavior, not their hypothetical answers. The limitation of the study is that it focuses on medicine, but the principles of «nudging» apply to other niches as well.
What This Means for Marketing Strategy
Marketers should reconsider the fear of brand «technological sophistication»:
- Personalization matters more than tools: People are hooked not by how you learned about their problem, but by how accurately you hit their context.
- AI is not a barrier: Clients show neither particular enthusiasm nor aversion to algorithms. Don’t spend ad budget on «selling» the fact that your service uses neural networks, unless it’s a unique selling proposition.
- Focus on benefit: The «personal risk» message worked as a trigger because it created a sense of care and concrete utility.
How to Apply This Data in Practice
To increase conversion through personalization, try the following steps:
- Shift the focus: Instead of «We use neural networks to select products» write «We analyzed your preferences to save you 2 hours of searching».
- Optimize headlines: Instead of «Our AI found the best discounts» use «We selected 3 products that perfectly fit your style — prices reduced».
- Focus on the solution: Instead of «The algorithm calculated the probability of a breakdown» write «Your car needs attention: check the brake system to avoid expensive repairs».
- Remove barriers: If you use complex technologies, explain not «how it works», but «what problem it solves for the user».
Frequently Asked Questions
Should You Hide AI Use in Mailings?
No, the study showed that mentioning algorithms does not affect conversion at all. You can write about it if it increases trust in data accuracy, but don’t expect sales growth from it.
Why Did Personalization Work Better?
It creates a «relevance effect». When a message concerns a person’s specific situation, it is perceived not as spam, but as a useful recommendation.
Conclusion
The main conclusion: don’t be afraid of technologies, but always put client benefit first. Want to check which approach will work on your audience? Calculate the effectiveness of your creatives with RGB Analytics and stop guessing what exactly hooks your buyers.
Source Nature Human Behaviour ↗
