Experiments, especially social ones, are now popular for the same reason reality television became popular.

They appeal to our voyeuristic instincts.

We get to compare ourselves with people who are dumped into situations that we may either wish we could be in, or are relieved that we are not.

For advertisers, they are a great option, because they are cheaper to stage and orchestrate than TV ads.

But in order for your audience to believe your experiment’s conclusions, it must be unbiased.

And so when you conduct one, you have to follow some scientific principles.

Your experiment should start with the framing of a question or a hypothesis that you want to test.

Moreover, every participant must go through the same procedure.

That doesn’t mean that you have to be serious, formal and dull.

Your approach can as lighthearted and entertaining as your subject and your brand’s tone of voice will allow.

You can also take some liberties in the telling of the story.

You may disclose to your viewers that it’s a branded experiment right up front. Or you may save it for a reveal right at the end.