Make your campaign experiments even more powerful

Back in July, we announced the rollout of campaign experiments to all Microsoft Advertising users globally. Today, this tool is more important than ever to campaign decision-making, especially as you consider how to increase your discoverability and maximize your agility. With people now spending more time searching for things online, you can adapt your marketing to show up in the moments that matter. Campaign experiments allows you to do this by testing out campaign changes without fully launching them across the entire campaign, allowing you to have full confidence in the decisions you make for your campaigns. And now we’ve made campaign experiments even better, with a new cookie-based split option.

What are the different options for setting up experiments?

As a reminder, once you have a clear hypothesis and goal in mind, visit the Experiments tab and you’ll see the tile below. You start by choosing the campaign you want to test, the name of the experiment, a start and end date, and then an experiment split. We recommend setting your experiment split at 50% for you to get enough volume quickly and make the comparisons more easily.


The Create an experiment page. The new cookie-based split option is now under Advanced options.

Finally, there’s the new Advanced options section where you can now select between two different options for setting up experiments: search-based and cookie-based. This is where you can now determine whether to show individual customers ads from the experiment and the original campaign randomly (search-based), or only from one or the other (cookie-based).

Search-based experiments

If you choose the search-based split option, every time customers search, they’ll randomly be shown either ads from your experiment or ads from your original campaign. This means that individual customers could potentially see ads from both sources if they search multiple times. This is the original split option that has been available since experiments first rolled out in July, and it’s great for testing certain features such as automated bidding tactics, since the comparison is on the bidding in each individual search auction.

Cookie-based experiments (new)

If you choose the new cookie-based split option, when individual customers search, we’ll show ads from either your experiment or your original campaign, and we use a cookie to ensure that, going forward, they’ll only see ads from this campaign. This is a great option for testing your creative and ad-copy, since you’re ensuring that an individual customer is only responding to one source or the other.

However, note that while you may potentially get more accurate data for some of these types of experiments, it may take longer to get statistically significant comparison data than if you did a search-based experiment.

What else is there to know?

Everything else about campaign experiments has remained the same, from monitoring your metrics via the scorecard to applying the changes you tested to your original campaign. To learn more, check out our July announcement for details and best practices on creating experiments, and see our help documentation: Discover the possibilities with experiments.

Let us know your thoughts

For any questions or feedback regarding campaign experiments, we encourage you to reach out to your Microsoft Advertising account manager or contact Support. You can also ping us on Twitter, or suggest a feature on the Microsoft Advertising Feature Suggestion Forum.
 

SEO Company in lucknow