Google Ads Budget Experiment: Playbook for Your First Test

Google Ads just added a new beta variable to the experiments area that most operators have been asking for. It is called Budget or bidding targets for Search campaigns, and it lets you A/B test different budgets, different target CPA amounts, and different target ROAS amounts against each other to find the spending level where a Search campaign actually performs best. The Google Ads budget experiment showed up in accounts on September 10, and it is worth setting up on at least one Search campaign this week.

If you have been raising or lowering budgets by feel every quarter, this is the tool that lets you replace guesswork with a real side-by-side test. Here is what the Google Ads budget experiment actually does, how to think about which campaigns to run it on, and the operator playbook to get one live this week.

What the Google Ads budget experiment actually tests

Start with the mechanics. Inside the Google Ads experiments area, you now see a new variable called Budget or bidding targets for Search campaigns. According to Search Engine Roundtable’s write-up, the experiment lets you test different budgets, different target CPA amounts, and different target ROAS amounts to find the optimal level of spending for a campaign. Google’s own description in the tool says the test is meant to identify the best spending level for the campaign you point it at.

The rollout also references A/B tests for different budgets and ROAS or CPA targets across multiple Search campaigns. That framing matters. This is not a single what-if calculator. It is an experiment structure that pits variants against each other and reports results the same way Google Ads experiments have always reported results, which means you keep your existing measurement approach and just add spend and target variables to what you already A/B test.

The experiment sits inside the same experiments area you use for AI Max tests and other Search campaign experiments today. If you already have a testing playbook in place, this new variable slots into that same routine. If you do not, this is a good reason to build one.

Why the Google Ads budget experiment matters right now

Two reasons this belongs on your radar this week even if you have plenty of other Google Ads work in flight.

First, this is the first time Google has given operators a native A/B way to test spend and bid targets together in Search. You could always guess. You could run a manual before-and-after. What you could not do was run a randomized test where budget or target was the only variable being changed. That is exactly what the new experiment enables. For any account where a client keeps asking whether the budget is right, or whether the target CPA should be lower or higher, you now have a defensible way to answer.

Second, this arrives right after the target-based bidding update from August. On August 17, 2026, Google updated its bidding systems so campaigns limited by budget and using Target CPA or Target ROAS now deliver more consistently toward the stated target. Google’s Google Marketing Live 2026 bidding and budgeting announcement laid out the direction earlier in the year. That change caused some campaigns that had been overachieving their targets to shift performance back toward the stated target. Advertisers who liked the old actual CPA had to lower their target to keep it. Now this new tool gives you a structured way to figure out what the right target actually is instead of adjusting by feel and hoping.

Put those two moves together and Google is saying the same thing twice. Bid targets and budgets are the levers that drive Search performance. Set them with data, not habit.

Which accounts should you set up a Google Ads budget experiment on this week

Not every account is a good first candidate. Use this quick screen to pick two or three where the payoff is highest.

Priority one: accounts where a Search campaign has been marked Limited by budget in the last 90 days. If Google has told you the campaign wants more room, testing a higher budget as one arm and holding the current budget as the other arm gives you clean data on what incremental spend actually buys you.

Priority two: accounts where the actual CPA or ROAS has drifted from the target after the August 17 bidding update. If your target CPA is set at 30 dollars and your actual CPA has moved from 22 to 28 since August, run the experiment with the current target as the control and a lower target as the treatment. That tells you whether the account can still hit the tighter number or whether you need to renegotiate the target with the client.

Priority three: accounts where you are about to raise or lower budget by more than 25 percent for a seasonal push. Instead of switching cold, run the change as an experiment first. A two-week test with a clear winner is worth more to the client than a full-month reallocation that you have to defend in the QBR.

Skip accounts where the Search campaign is brand-only, where daily budget is under 20 dollars, or where conversion volume is under 30 per month. The experiment needs enough signal to reach a result, and thin campaigns will just chew through your test window without giving you a clear read.

How to set up your first Google Ads budget experiment

Block one hour on your calendar this week. Here is the sequence.

Step one: pick the winning campaign from your screening list above. Note its current budget, current bidding strategy, current target CPA or target ROAS, current actual CPA or ROAS from the last 30 days, and current conversion volume. Screenshot the campaign settings so you have a clean baseline.

Step two: open the experiments area in the campaign and select the new Budget or bidding targets for Search campaigns variable. Pick whether you are testing budget, target CPA, or target ROAS. Do not test more than one at a time in the first experiment. You want to be able to attribute the result to a single variable.

Step three: set the experiment split at 50 percent. Do not do 10-90 splits on your first test. You want enough traffic on both arms to reach significance in two weeks.

Step four: define the treatment value. For a budget test, try plus 25 percent or minus 25 percent. For a target CPA test, try minus 15 percent to see if you can hold volume at a tighter target. For a target ROAS test, try plus 10 percent to see if you can push efficiency without collapsing volume.

Step five: schedule the experiment for 14 days. Note the end date in your project tool and set a reminder to review results the day after. Do not extend on the fly. If the test is inconclusive at 14 days, run a follow-up test with a bigger treatment value, not a longer window.

Step six: document the win condition before you launch. Write down the exact metric that will decide the winner, the minimum lift you need to justify rolling out, and the fallback plan if the treatment loses. This turns the experiment from a random test into a decision-ready deliverable for the client.

How the Google Ads budget experiment fits with everything else this quarter

September is already dense for Google Ads operators. The AI Max auto-upgrade is landing on Search campaigns. Shopping is running through policy consolidation. Search accounts globally have the language targeting change to sort out. This test is one more thing on that list, but it is also the tool that helps you defend budget and target decisions once the dust settles.

Sequence the work. Do the AI Max upgrade review first because that changes campaign structure. Do the language targeting audit second because that changes traffic composition. Then set up the budget test third, once the campaign is in its new steady state and the data you collect during the test will actually reflect the setup you plan to keep. Running a budget test on a campaign that is about to structurally change is a waste of a two-week window.

If you also manage Shopping campaigns, the same discipline applies to the Standard Shopping migration and the AI Max for Shopping rollout. Get the structural changes stable first, then run the spend and target experiments on top of that stable base.

Making the Google Ads budget experiment part of your monthly rhythm

The teams that get the most out of Google Ads experimentation do not treat it as a one-off. They set a standing rule that every managed account runs at least one experiment per month, and one of those every quarter is a spend or target test. That gives you 12 defensible spend decisions per account per year instead of two or three big annual reallocations that live and die on gut.

Add the Google Ads budget experiment to your monthly checklist next to your regular AI Max, RSA, and audience tests. Log every experiment in the same tracker so you can look back at the end of the quarter and see how many hypotheses you tested, how many won, and what the aggregate lift was. Clients love that data. It changes the meeting from what should we do next to here is what we learned this quarter and here is what we are testing next.

If you want a second set of eyes on which of your accounts are the best first candidates for the Google Ads budget experiment, or you want help writing the test brief so the win condition is airtight, book a free consultation and we will walk through it together. Let’s Grow!