Designing Successful Experiments

Incrementality experiments are a powerful tool to determine the causal impact of your marketing campaigns, but they require careful experiment configuration. There are several actions you can take to help ensure a successful experiment:

1. Ensure that you have given enough time for your experiment to be successful

Many digital marketing strategies can take quite a while to optimize, and Gamesight recommends a minimum experiment time of at least 2 weeks.

2. Verify you have enough data to create a baseline

As a rule of thumb, you should have at least a month of data in at least 20 regions prior to running your experiment

3. Ensure your campaign has enough budget to be successful

Incrementality experiments require enough spend to see a signal over the noise of your baseline. If you're struggling to see lift, consider increasing your spend

4. Consider testing a single strategy at a time

Unlike attribution, incrementality cannot break down your performance results between multiple marketing strategies

5. Be sure to eliminate extraneous variables from your experiment

While incrementality tests can be successful running alongside global campaigns, any marketing in your test and control regions is likely to affect your results. You should do your best to ensure test and control regions are not preferentially targeted by any ongoing marketing strategy.

📘

Still have questions about Incrementality?

If you're unsure about your experiment, or are having trouble with your results, please reach out to your Gamesight representative or contact [email protected]!