A\B Tests
Design & Analysis
To consult the statistician after an experiment is finished
is often merely to ask him to conduct a post mortem examination.
He can perhaps say what the experiment died of.
Ronald Fisher
Throughout our years in the industry, we observed many experiments that companies, large or small conducted.
The amount of methodological errors that were made is astounding!
When not designed carefully, methodological errors may arise in any one of the phases of an experiment, from data collection, sampling users\test-subjects, group allocation and randomization, and all the way to analysis of data, we’ve seen it all.
Of course any mistake that is made makes the data from the experiment useless and one should not make any decisions based on it, making it a waste of time and money.
With us, no experiment is lost. We are familiar with the appropriate methodologies for any type of data and business problem.
We provide a tailor-made experiment to your needs, which more often than not, extends further than the basic control\treatment-group comparison.
There are much more sophisticated ways of conducting experiments!
Ones that provide a deeper understanding and higher resolution of the data.