What are they?
Statistical experiments are tests carried out in Pareto to plot, with statistical confidence, a learning curve about the brand and its audience.
#WeAreAgainst
We don't agree with small experiments without the correct conception of the hypothesis. That's why we've created the video below to quickly summarize, in 5 slides, how to work with experiments in Pareto!
The Hypothesis
The hypothesis is critical to creating an experiment. Pareto has created a hack to facilitate its creation. Complete it below:
IF my ad varies ____________
THEN I hope to optimize my results WHY _______________________
The why in the sentence above carries with it THE RATIONALE of the test. Let's take an example:
Since my competitors don't work on the color of the beds in their ads, I'm going to try it out!
IF my ad has the color of the shirt in the title
THEN I expect to optimize my click-through rate because my competitors are generic and don't go that deep.
How to work with Pareto experiments
Every quarter you can send us a new batch of experiments that you want to run. Remember that we always test an original versus just ONE VARIATION! In other words, one test at a time.
Formulate your hypotheses and pass them on to our team.
Learning curve
Nowadays we test everything (out there) and come up with nothing. Let's work on a few tests, but build a learning curve to better understand your consumer's behavior and the triggers to activate it!
Multitest Vs Test
To know that FREE SHIPPING has improved your CTR, it's important that the only change you make to your ad is to insert the term FREE SHIPPING. If you completely rewrite the ad, you are already testing the whole new ad and not just the term FREE SHIPPING.
Original Ad
Pareto sale with 20% off
Announcement Variation
Pareto sale with 20% off and free shipping
Multitest ad
Come and shop at Pareto. Everything with free shipping.
There's no problem with doing multi-tests instead of single tests. However, be aware that you won't conclude that FREE SHIPPING is better for your business, but rather that the entire new ad performed better.
