How To Use Campaign Experiments To Optimise Your Google Grant Account

Testing and experimenting are a major part not only in Google Ads but in any digital marketing campaign. There is no one-size-fits-all strategy and it is worth the effort to investigate what makes most sense in your individual circumstances. 

Especially in Google Grant accounts – which come with some limitations – fine-tuning your campaigns, ads and keywords can help improve quality and thus overall campaign performance.

What are Google Ads campaign experiments?

A campaign experiment is an A/B test to test two slightly different versions of the same campaign against each other. Instead of modifying an existing campaign based on a hypothesis and just hoping for the best, an experiment is a low-risk way of trying out minor campaign modifications before fully committing to the change.

During the setup process, a copy of the original campaign is created and then traffic and budget are split equally between the campaigns allowing for easy evaluation and statistically significant data.

Once enough data is available and you identified the better-performing campaign version, you can simply end the experiment and keep either the original campaign or the new experiment campaign.

What can you test?

While you can vary nearly any setting in your campaign setup, the following are the most common use cases for campaign experiments in Search campaigns:

• Ad copy: Test different versions of ad messaging, for example, a more casual tone of voice against a more formal one.

• Keyword match types: The most common keyword match type these days is broad match but in some cases, phrase or exact match may be a better option to get more relevant traffic or limit unwanted searches.

• Bid strategies: This may be limited depending on your organisation type and goals and how visitors can interact with your website. For example, it could be a simple ‘max conversions’ vs ‘max conversion value’ test if you want to increase sales or donations or you want to find out whether including a Target CPA in your bid strategy gives you better results.
  

• Landing pages: This is one of the most common variables to test. You can have the best campaign setup but if your landing page isn’t convincing, you lose visitors instantly. It is a great way to find out what your audience best responds to and these insights can help not only get better campaign results but improve your entire website.

How to set up a campaign experiment

Setting up an experiment is an easy process and Google guides you through it step-by-step.

Start by going to the Campaigns tab in your Google Ads account, select “Experiments” and click “Create an experiment”.

Choose “Optimise text ads” if you want to test another version of your ad copy. For anything else, choose “Custom experiment”.

Note that Google Grant accounts allow Search campaigns only, so the other options presented on this screen – Video and Performance Max experiments – wouldn’t work unless you have a standard Google Ads account.

Give your experiment a descriptive title and select the base campaign you want to create the experiment for, then click “save”.

Your experiment campaign is now set up and you can start making the change you want to test. Simply edit the campaign as you would do in a normal campaign by going to the ad copy, keywords or settings section and making the respective change. Once you made the changes, click “Schedule” and your experiment will start.

The setup guide isn’t very obvious when doing this for the first time and can be overlooked easily leaving you wondering whether you’re still in setup mode. Keep an eye on the top of the page to see where you are in the process and where to find the “Schedule” button:

Best practices for campaign experiments

While the experiment setup is pretty straightforward, here are some tips to get the best out of your experiment.

• Don’t set an end date or duration for the experiment but keep it running until you have enough valid data to make a decision. Depending on how much traffic your campaign usually gets, this may take a few weeks or months. Keep in mind that traffic is split 50/50 between the original and the experiment, so if your base campaign usually gets low traffic, getting enough data can take some time.

• Your experiment campaign will appear in your campaign overview next to all other campaigns. If you go to the experiment section, you will see the exact performance comparison of the original and the test campaign within the time period of the experiment. You can compare the metrics that matter to you most and see if the results are conclusive or if there is not enough data yet.

• For effective testing, make only one modification in your experiment campaign. Testing multiple variants such as a different landing page combined with a different bid strategy won’t give you clear insights. You would never know which of your changes (landing page or bid strategy) was the main driver of a change in performance and once again end up guessing.

Conclusion

Campaign experiments are an easy, low-risk way to test minor changes in your Google Search campaign setup. They can provide great insights into what works for your individual situation and play an important role in the long-term optimisation of your campaigns to achieve the best possible results.