Google Ads Conversion Lift: Measure Incremental Sales

Google Ads conversion lift is a measurement solution that tells you whether your ads are actually driving new conversions, or if people would have converted anyway. It works by running a controlled experiment: a portion of your eligible audience is held back from seeing your ads (the control group), while the rest sees them as normal (the exposed group). By comparing the conversion rates of these two groups, you get an incremental lift percentage and the incremental conversions your ads truly generated. This guide walks you through setting up and running a conversion lift study in Google Ads, from prerequisites to interpreting the results.

What You’ll Need

• A Google Ads account with an active campaign (Search, Display, YouTube, or Discovery) that has been running for at least a few weeks and has sufficient budget.
• Conversion tracking set up and working correctly (at least one primary conversion action, such as a purchase or sign-up).
• Enough historical data: Google recommends at least a few hundred conversions per month for reliable results.
• Access to the Google Ads account with administrative permissions.
• Patience: the study typically runs for 4–6 weeks to collect statistically significant data.

Step-by-Step

1. Verify your conversion tracking. Go to Tools > Measurement > Conversions. Ensure your conversion actions are active and recording data. Check that your global site tag or Google Tag Manager is firing correctly on your conversion pages (e.g., order confirmation).
2. Check campaign eligibility. Your campaign must meet Google’s minimum thresholds: usually at least 30 conversions in the last 30 days, and a daily budget that can support the study. You can check eligibility directly in the Google Ads interface when you set up the experiment.
3. Navigate to Experiments. In the left menu, click on “Campaigns”, then “Experiments”. If you don’t see it, click “More” at the bottom of the menu.
4. Create a new experiment. Click the “+” button and select “Conversion lift” from the list of experiment types.
5. Select the campaign. Choose the campaign you want to test. The tool will show you an estimate of how long the study will need to run based on your current conversion volume.
6. Set your control group size. You’ll be asked what percentage of your eligible audience should be held back as the control group. A common default is 50%, but Google will suggest a percentage based on your traffic. A larger control group gives more statistical power but reduces the number of people who see your ads during the test.
7. Review and launch. Check the settings, then click “Start experiment”. Google will automatically split your audience and begin collecting data. You don’t need to create a separate campaign; the system handles the holdout.
8. Monitor the experiment. While the study runs, you can check its status in the Experiments page. You’ll see a progress bar and estimated time to completion. Avoid making major changes to the campaign (like changing bids or budgets) during the test, as that can invalidate the results.
9. Read the results. When the experiment finishes, you’ll see a report with the following key metrics: Conversion Lift (the percentage increase in conversions due to your ads), Incremental Conversions (the actual number of extra conversions your ads drove), and a Confidence Interval (the range in which the true lift likely falls).
10. Apply learnings. If the lift is positive and statistically significant, your ads are working. If the lift is near zero or negative, your ads may not be adding incremental value, and you might need to rethink your targeting or creative.

Common Mistakes

• Ending the experiment early. If you stop the test before it reaches statistical significance, the results are meaningless. Wait for Google to indicate the study is complete.
• Changing campaign settings mid-test. Altering bids, budgets, or audiences during the experiment can skew the results because the control and exposed groups may no longer be comparable.
• Ignoring the confidence interval. A lift of 20% might sound great, but if the confidence interval is -5% to 45%, the true lift could be zero. Always check the interval before making decisions.
• Using conversion lift on low-volume campaigns. If you don’t have enough conversions, the test will take an extremely long time or may not be possible. Check eligibility first.
• Not having a clear primary conversion. If you track multiple conversion actions (e.g., newsletter signups and purchases), the study uses your primary one. Make sure it’s the one that matters most to your business.

Tips

• Use a larger control group for faster results. A 50% holdout gives you more statistical power and can shorten the study duration, but it means half your audience won’t see your ads during the test. Balance this with your need for traffic.
• Run the test on your highest-spend campaigns. You’ll get more reliable data faster when the campaign has significant volume.
• Combine with other experiments. You can run a conversion lift study alongside a draft experiment (like a bid strategy test) to get a fuller picture of performance.
• Be patient. The default study length is often 4–6 weeks, but it can be longer for low-volume campaigns. Don’t rush it.
• Use the results to inform budget allocation. If one campaign shows high incremental lift, consider shifting more budget to it. If another shows zero lift, you might reduce spend.

Troubleshooting

• Experiment is taking too long. If the estimated completion date is far in the future, you can increase the control group size (if the experiment hasn’t started yet) or add more conversions by improving your conversion tracking. Once started, you cannot change the control size.
• No data appears after a few days. Check that your conversion tracking is still firing. Also, ensure your campaign is still active and getting impressions. If the campaign has a very low volume, the experiment may take longer to show data.
• Results show “Insufficient data”. This means the experiment couldn’t reach statistical significance. You may need to run it longer or with a larger control group. Unfortunately, you’ll need to start a new experiment.
• Conversion lift is negative. This can happen if your ads are attracting people who wouldn’t have converted otherwise, but it’s more often due to random chance. Check the confidence interval – if it includes zero, the negative lift is not statistically significant.
• You accidentally paused the campaign. If you pause the campaign during the experiment, the study will be paused too. Resume the campaign as soon as possible to avoid invalidating the test.

Frequently Asked Questions

How long does a conversion lift study take?

Typically 4–6 weeks, but it depends on your conversion volume. Google Ads will give you an estimate before you start. Low-volume campaigns may need longer.

What is a good conversion lift percentage?

There’s no universal benchmark – it varies by industry, campaign type, and baseline conversion rate. What matters is whether the lift is statistically significant and positive. Even a 5% lift can be valuable if you have high volume.

Can I use conversion lift with a Smart Bidding campaign?

Yes, conversion lift works with Smart Bidding campaigns. The experiment holds out a portion of your audience, and the bidding algorithm adapts to the exposed group. This is actually a great way to measure the true incrementality of your Smart Bidding strategy.

Does conversion lift work for brand campaigns?

Yes, but be aware that brand campaigns often have high baseline conversion rates, so the incremental lift may be small. The study will still tell you if your brand ads are driving additional conversions or just capturing people who would have found you anyway.

What’s the difference between conversion lift and a draft experiment?

A draft experiment tests a change to your campaign (like a new bid strategy or ad copy) by splitting traffic between the original and the draft. Conversion lift tests the overall effectiveness of your ads by comparing against a no-ad control group.

Can I run multiple conversion lift studies at the same time?

Yes, but each study requires a control group, which reduces the audience available for other tests. Running too many simultaneously can slow down all of them and reduce statistical power. It’s best to prioritize your most important campaigns.

Related

• Doxxed Staff or Volunteers: What to Do
• App Corrupted After Mac Cleanup? Fix It
• Elden Ring Open World & Difficulty Explained