"In our sample, an advertiser that ran 15 experiments (versus none) in a given year sees about a 30% higher ad performance that year."
- Julian Runge, Facebook's marketing science research group
Feature | Meta's Native Testing Tools | Creative Analytics Tools | A/B Testing Tools | Deciding Data |
---|---|---|---|---|
Identifies winning ads | ||||
Statistically valid | ||||
Well formatted results for sharing | ||||
Makes good experiment design easy | ||||
Identifies creative features across ads and tests | ||||
Connects creative features to performance | ||||
Translates performance data into undestandable creative and company lessons |
Upload ads - we can test images, videos, text, and targeting options.
We suggest optimal testing budgets based on your content and CPO. We maximize learning at any spending level.
Multiple small-scale tests are launched, cleverly designed to build a statistically valid and comprehensive picture.
Tests run in your ad account, ensuring compatibility with existing attribution and data systems. It works like your other ads. You maintain full control.
Watch results unfold as experiments run. We identify features across your ads that may be driving performance and provide clear probabilities for informed decision-making.
Knowing why ads perform well helps you identify what to run next and builds a shared understanding of your customers' needs.
Facebook provides essential A/B testing capabilities, Deciding Data builds on that. We provide test orchestration and more advanced analysis. Our analysis combines the results for many tests and works to uncover the reasons behind ad performance. So you can make more informed decisions and iterate more effectively.
Sadly, No. To understand ad performance, and not just report on it, you need randomized tests setup in particular ways. Without randomized experiments, influences like trends or ad fatigue can strongly bias performance.
We are more Scientist, less Accountant. We don't just want to know what is performing best, we want to know why and what we can do to improve future ads.
Beyond tracking to understanding why
While most creative analytics tools show where users engage with ads, Deciding Data identifies why certain ads outperform others. We identify specific creative features (like "Body Focus" or "NYC Skyline") that drive performance.
We're Scientifically Sound
We use randomized tests to learn the truth, avoiding bias from trends or campaign structure differences that typical analytics miss.
From Question to Creative Direction
Think of us as scientists that translate results to creative lessons. Where creative analtyics tools put together reporting, we give you the creative recipe for success.
Our platform:
- Spots patterns across multiple creative variations
- Identifies actual creative elements and styles in your ads
- Translates performance data into specific creative direction (e.g., "NYC Aspirational aesthetics outperformed Product Focused approaches")
Creates Shared Creative Strategy
Deciding Data helps you focus resources on elements that work. This builds a consistent data-backed creative strategy that is understood across teams and evolves with your audience.
"Statistically significant" means: If we assume these two ads are the same, it's very difficult to explain the difference in performance with random variation alone.
Waiting for "significance" can be wise when change is risky or hard. But with ads, change is easy and often needed. We have to pick something.
Instead of just yes/no answers, we use probabilities. We'll tell you "there's a 77% chance an image with this focus produces the highest ROAS." This helps you:
- Know what to choose
- Understand the strength of evidence in the data
- Incorporate other information in your decisions
Want to dive deeper? Our founder wrote about balancing inaccuracy and inaction here.