Google Ads Experiment for Beginners – Google ads experiment for beginners, With Google Ads, it’s common to question how a change in your settings or assets will affect your campaign. Is there a reason why you’ve been reluctant to change your settings because you can’t afford to put all of your eggs in one basket? Maybe you’re not sure if the adjustment you’re proposing will make a significant impact. You may already be convinced, but you need help persuading your client or management.

Google Ads Experiment for Beginners

Any time you’re making a major shift in campaign approach, it’s a good idea to have some statistics to back up or disprove your claim. Initiate campaign tests.
They’re new to you, aren’t they? Break it down for us.

It’s not hard to figure out what Google ads experiment for beginners

Consider them a tool for A/B testing.

It’s important to conduct trials in order to see how modifications will affect your campaign before making them permanent. Experimentation yields real-world user performance data from which to draw conclusions regarding the best course of action.

How does that work? We’ll get into the details below, but essentially, experiments allow you to run an alternate version of your campaign in tandem with your original, measuring and comparing performance against your existing setup along the way.

What’s the point of conducting marketing experiments?

Why not make a modification and see how it affects your results?

That’s a strategy you may try. In the end, is that the best way to go about it? I don’t think so.

In order for your Google Ads campaigns to succeed, they must be integrated with your whole marketing strategy. It doesn’t matter what business or product you’re in; external variables will always have an impact on your performance.

Testing changes in the same context as the original campaign is possible by doing campaign experiments. Seasonality, current events, and changes in your competitor’s strategy can’t explain the disparities in performance that you’ve noticed.
Experiments are your best buddy if you care about comparing like with like data.

Experiments can be used to investigate a variety of topics.

In general, experiments can be used to test most campaign components, however there are a few caveats to be aware of:

Campaign types

  • Only search and display campaigns can benefit from experiments.
    No app, video, discovery or commerce efforts are allowed to use them
  • Budget
    • To fund experiments, a portion of the original campaign’s budget is allocated to each one. There is no way to compare two budgets side by side.
      Ineligible are campaigns with shared budgets – Just change the campaign’s budget to an individual one if you wish to do a test.
  • Audiences
    • In order to conduct studies with a cookie-based split, audience lists must contain at least 10,000 individuals.
  • Ad customizers
    • It is not possible to use the following ad customisers togetherwith experiments:
      • Target Campaign
      • Target Ad Group
  • Each campaign is limited to a single active experiment.
    Up to five tests can be scheduled in advance.

Campaign experiments should be used when?

If you’re unsure of the impact of making significant changes to a search or display campaign, we recommend running an experiment first. For testing campaign adjustments, campaign experiments are by far the best method. Experimentation is the only way to find out for sure whether or not a change was successful.
Campaign experiments are frequently used to test the following things:

  • Different bidding strategies
  • Device bid adjustments
  • Alternate landing pages
  • Ad schedule
  • Audience targeting/bid adjustments for search campaigns
  • Ad copy changes

Using tests to test ad wording and landing pages isn’t the only way to do this.

Only one campaign at a time

Testing ad text is as simple as creating many ads within ad groups, each with a different creative, copy, or end URL. winning ads. 

Do not optimize: Rotate advertising endlessly is a good option if you want your new ads to rotate evenly with the originals. Don’t forget to modify your performance optimizations back once you have the data you require if you did so before the test. In order to learn more about ad rotation, please see our Google Ads settings page.

For multiple campaigns

Instead of tests, use Google’s Ad Variations function if you’re running different campaigns with the same ad wording or landing page.
We’ve covered the “why,” “what,” and “when” of the situation (whenever you want to test the impact of a major change to a campaign). That should get your testing mojo flowing in the direction you want it to go.

Campaign experiments best practices

The most important thing to follow while doing experiments is to just test one modification at a time, with the rest of the campaign remaining unchanged. While it’s tempting to handle all of your campaign optimization improvements at once, this isn’t how campaign optimization works.

You can experiment with several changes to a campaign, but you’ll be left wondering which of those changes had the most impact…. That’s not something we recommend.
Take one step at a time, and then move on to the next until you get it right.
Are you all set to put your skills to the test? Come on, let’s see how this works.

Experiments can be created in Google Ads by following these steps:

In order to begin testing, you’ve decided to set up a Google Ads campaign. Great! Prior to making any changes, Google will ask you to create a “draft” of your changes.

See also  Determining the Best Keywords Google Adwords PPC Advertising

What is a campaign draft?

 Your campaign can be edited in a draft form without affecting the live campaign. In order to test a new idea, you can make revisions to your draft and then implement them in your campaign.

Creating a campaign from an existing draft

Navigate to the Drafts & Experiments drop-down menu in the Google Ads online interface and select the “Campaign drafts” option. Open your draft by clicking on its name:

Proceed to the next step, which is to set up your new experiment.
Need a rough draft created? We’ll show you how it works.

Creating a campaign draft

As a first step, make sure that you’re using Google Ads’ online version. It is not possible to publish drafts in Ads Editor without setting them to “enabled” or “paused.”

“Experiments” can be found under the “Drafts & Experiments” menu of the campaign you want to experiment with. Make sure “Campaign drafts” is selected from the drop-down menu.

If this is your first draft, you’ll have two sparkling options to choose from!

Your draft will be given a name. Name it in the same way you would any other campaign—descriptively and in accordance with your account’s name guidelines. We like to name our drafts after the experiment we’ll be conducting, but this isn’t required. Make a plan that works for you and follow it religiously!

If you’re looking for some inspiration, we recommend checking out:
Campaign Name: Thing Being Tested Against Control
You can see campaign and experiment data side by side when sorting All Campaigns alphabetically after naming the experiment as such while you’re looking at the All Campaigns tab. It’s also a watertight mechanism, so your Google Ads account may be managed by many persons without confusion.

For example:

CPA vs. CPC: Which Should You Aim For When Listening to Cat Purrs?
You can also include a description in addition to the name. If you’d like to provide further information, such as the CPA that you’re testing, or the ad copy that you’re testing, you may do so here.

Hit save.

Your new draft campaign’s Ad Groups tab will appear. Any similarities? It’s supposed to be like this! A draft of your original campaign has been created. When you’re in the draft, you’ll be able to tell because the top of your screen will state so and the right side of your screen will show the tab you’re in.

Changing the element you’re testing is as simple as navigating from this point onward.

The editing process is the same for your draft as it is for a standard campaign, so you’re comfortable with it. Make changes to your ad groups, bid strategy, etc. under the settings.

If you’re interested, you can see how experiment traffic differs from your site’s original campaign traffic to get some food for thought. Consider changing your tracking parameters if necessary for the experiment.

What’s the matter? Here’s a link to our guide on how to track your campaigns.

If you aren’t ready to do an experiment at this time, you may exit the page without completing it. You don’t need to take any further action to save your draft, and it will be accessible from the Campaign Drafts tab.

Setting up your new experiment

The experimentation phase begins once you’ve completed the necessary revisions to your manuscript. You can apply by clicking the APPLY button on top of the navigation bar.

To help you decide, you’ll be given two options:

Adapt the original campaign instead of selecting this one; you’re not here for that.
Choose this choice and then click APPLY again to conduct an experiment.
What follows is an extensive list of parameters that you can use to customize your experiment.

  • Name
    • If you need help coming up with a name for your experiment, see our suggestions in the draft area above.
      No two experiments can have the same name, paused or activated. When an experiment is deleted, the name is made available for future usage.
  • Description
    • Additional information can be entered in this optional parameter – see our note about descriptions above for more details.
  • Start date
    • As a general rule, we recommend that you leave the experiment start date as it is, or set it for a future date. Aside from saving time by not launching your experiment until the next day, this also allows your ad campaigns a chance to be reviewed and approved.
    • Although the ad copy hasn’t been altered, it still needs to be approved by Facebook’s team.
      Campaign experiments that have been successfully developed but are not currently running will show a “Scheduled” status.

When does the project end?

Your experiment will come to an end on this date.
Choose this option if you want your experiment to continue for as long as you want it to continue. It’s up to you when you want to terminate it. If you want your experiment to end on a specific date, specify a date in the future. At the very least, give yourself 30 days to acquire as much data as possible.

It doesn’t matter when you set an end date for your experiment; you can always return and manually close it. If you’d like, you can make it longer by modifying the expiration date.
Experiment with a split-second delay.

See also  Google Keyword Planner Tutorial (Guide)

“The amount of your campaign’s budget allotted to your experiment, and the percentage of auctions your experiment is eligible to participate in,” Google explains. For all intents and purposes, you’re deciding how much of the campaign’s money should be allocated to an experiment and how much to a conventional campaign.
In order to test both versions of your campaign equally, we recommend leaving it at 50% by default.

Be aware that even with a 50/50 split, KPIs like impression share and budget may not be distributed evenly. It’s very typical for this to happen, as you’re testing for performance improvements.

Experiment with different configurations for the split (search campaigns only)
Allows you control how your users will be presented with the results of your experiment and original campaign. To see this option, open the Advanced settings menu.

Your experiment split can be done in two ways:

When a user performs a search, Google will randomly choose one of your experiments or your original campaign to show up. This implies that if a person does a search that activates your campaign more than once, they could receive both versions.

Using a cookie-based approach, a user will only see the experiment or the original campaign, no matter how many times they make a relevant search, regardless of how many times they search for the same term. Because the consumer sees the same ad/landing page each time, this is a great way to test creativity because it minimizes the impact of other influences.

To make use of cookie-based split testing, you must have at least 10,000 users in each of your campaign’s target audiences.

Only search campaigns provide you the option of customizing your split; all display campaigns rely on cookie-based splits.

Changing the experiment

Don’t mess with your experiment once it’s running!
What’s the harm in trying? Your performance may likely be affected by any adjustments you make When looking at performance data before and after a modification is made, it can rapidly become a muddle, even though you can vary the date range. Don’t put yourself through it.

Unavoidable circumstances may force you to make a shift, and that is understandable. Make sure you remember these points if that’s the case.

Any modifications you make to your original campaign won’t be reflected in your experiment unless you manually alter it yourself.

Has there been an error in the original? The experiment should be paused as well, or the data will be distorted. Consider the importance of ensuring that all other aspects of your experiment are consistent.
Once an experiment goes live, it will appear in the Google Ads Editor section of the platform.
What you’re testing may be affected more by some modifications than others.

Make sure you know why you’re making the modification in the first place and how it could effect both the experiment and your control before deciding to implement it.

You can extend an experiment to obtain more data, for example, if you’re testing eCPC bidding against maximum conversion bidding and you decide to do so after two weeks. There’s no need to worry; modifying the test’s end date will have no effect on how well your campaigns work (though do consider seasonal changes before you extend a test for too long).

Is it possible to acquire more funding for an experiment halfway through? Be aware. Adjusting budgets for campaigns with maximize conversion bidding has led to substantial variations in cost per acquisition (CPA). The current experiment should be stopped and a new one should be started with the new budget.
You can do whatever you want with Google.

Be careful while making modifications to your experiment; Google will not stop you from modifying the very thing you’re trying to find out about.

Experimenting with target CPA bidding, for example, if you accidentally change your experiment campaign’s setting to maximize conversions in the middle of your test (of course), there will be no warning and no “are you sure you mean to do that?” messaging and the change will be implemented.



It’s done! Your first campaign experiment is in motion now. Relax and let the test take care of itself now.Of course, you’ll want to keep an eye on it throughout the process. 

The good news is that Google Ads makes it very simple to do so! You may see the results of your experiment in a few different locations.

In the Campaign experiments tab, you’ll find If your experiment is currently running, you’ll be able to see this information here. It’s easy to get to the actual experiment by clicking on its name.
All campaigns can be found under the Campaigns tab: Isn’t that view spectacular? This is it. There are separate rows for experiments and other campaigns. A beaker icon instead of a green circle indicates that a campaign is in an experimental state.

The totals at the bottom include all of the experiment data.
All campaigns’ ad groups can be seen in the Ad Groups tab. In the Ad Groups page, you can see all of your experiment ad groups alongside your standard campaigns.

See also  How to Advertise on Google Ads for Beginners

Alternatively, you can monitor the results of your experiment to see how it compares to a typical campaign: As soon as you select your experiment from the left-hand menu, you’ll be able to dive right in.
You can look at the history of ad groups, keywords, devices, and so on in this view. When it comes to comparing your experiment to the original, this is where things truly become interesting…
In most experiment tabs, you’ll see a panel above the typical graph and table view that highlights specific metrics. At any given time, the panel displays five different metrics, but you can switch between them by selecting one from the following list:

  • Conversions
  • Clicks
  • Cost
  • Conv. value
  • Conv. value / cost
  • CTR
  • Avg. CPC
  • Impr.
  • All conv.
  • Conv. rate
  • Cost / conv.
  • View-through conv.
  • Value / conv.

When you hover over these metrics, you can see how your experiment compares to the original. A statistically significant difference can also be determined if adequate data is supplied.

This is quite cool, don’t you think?

Pro Tip: The comparison of performance will reveal data for dates when the experiment is/was active. But the graph and table below are based on the reporting period you have selected.
As long as you’re looking for data from that time period, you can simply click on the experiment date range on the right side of your performance comparison panel to modify your dates.

A date range that is consistent with the experiment will now appear in your reporting.
In order to have a successful campaign, you need to keep track of your experiment at all times. Keep an eye out for updates on your experiment’s progress now that you know where to look. Keep an eye out for significant gains or losses in performance throughout the course of the trial (preferably not the latter).


It is possible to end a campaign experiment manually or automatically at a pre-determined date and time. To make it simple to implement or not, Google Ads offers a variety of options for making modifications to your ad.


Prematurely terminating an investigation

You can discontinue your experiment at any time, even if you’ve chosen a certain end date in the past. The data you need to make an informed decision is at your fingertips on day 43 of your 60-day challenge. Experimentation can be terminated at any time by you.
Go to a tab inside an experiment campaign where the performance comparison panel may be accessed, and click on the performance comparison panel. You can see all of your selections in the upper right corner of the page.

Click END NOW to put an end to the experiment if your initial campaign is the clear winner. Once an experiment is ended, it cannot be restarted. You will be asked to confirm that you actually want to do this.

Select CANCEL if you change your mind and want to continue this project. Click on END if you’re done playing. Your experiment will be completed, and your original campaign will continue as normal.

To make the change official, select APPLY after your test results are in. This is the end of your experiment, and you will be given two options:

Once you’ve made your choice, click Apply again. Time to switch it up! It looks like the change you were experimenting with worked. In some ways, it resembles magic, but it’s grounded in fact.
Applying the experiment’s status. You have the option to remove it at a later time without harming the new (or updated) version of the original campaign. Your experiment and all of its data will be permanently deleted in the same way that a conventional campaign is.

It came to an end.

You can still access your experiment from the Campaign experiments page of your original campaign after the end date you’ve selected has passed. To assist you evaluate the results of your experiment and make a choice on whether or not to implement the modification you’ve been testing, the performance comparison panel will remain available.
In most ways, the view will be the same as what you’re used to from following the experiment as it progressed:

No more action is required if your experiment was a failure and you wish to remain with your original campaign. It is possible to remove the experiment by clicking on REMOVE. Remember that deleting an experiment permanently removes it.
Pro Tip: If you don’t delete the experiment when it’s finished, you’ll still be able to access it. Using historical trials is possible, but we don’t encourage it because the outcomes are likely to be drastically different. Do another test if you’re unsure.
To update your campaign if you’re satisfied with the results of your trial, you can do it right here.


Click APPLY once more when you’ve found the solution that works best for you.

The experiment’s state will change to Applied when you accomplish this. It won’t effect the new (or updated) version of the original campaign if you decide to remove it at a later time.