Control testing is a type of A/B testing conducted to ensure that a test is valid and not a result of bad practices. The control segment allows researchers to compare the results against a baseline during the same time period as the tested changes, which helps account for seasonality and other anomalies.
Normally the control segment is the A group, while the B segment is the test.
Without control testing, the results from a given test are hard to prove. For example, if you sent out a coupon for 20% off and revenue skyrocketed the following month. How do you know it was the coupon with 100% certainty? Could it be seasonality, and the coupon was hurting order value instead of helping? The only way to know is to compare sales to a control segment that wasn't given the coupon. The difference between the two numbers is called lift.
Establishing a control group is essential in any web analytics A/B test.
Optinmonster is not built for control testing, and as a matter of fact, customer support was unable to help us set up a control group whatsoever.
To jerry-rig the tool, I had to add a 'blank' popup, which is a normal popup except everything is set to transparent.
We were able to track who was shown the blank ad within Google Analytics. Turing Optinmonster into a powerful A/B testing tool.
In a cart, most customers leave and never return. According to the Baymard Institute, the abandonment rate was 69.6% in 2022.
Cart abandonment is normal behavior; customers need additional time to make a purchase. For the site in question, customers averaged 183 days before checking out. Many users left on the first and second steps.
Optinmonster is a mail popup tool that shows users' interstitials depending on several factors, such as page URL, time on site, or other actions.
We can increase sales by showing a targeted audience an interstitial pop-up upon exit intent.
Optinmonster has extensive multivariate testing options, and the interstitials Optinmonster produces trigger events within analytics, making tracking engagement extremely easy.
However, within Optinmonster, we cannot add control groups. Optinmonster is not built for control testing, and as a matter of fact, customer support was unable to help us set up a control.
To get the interstitials to display a proper control, we used the same conditions that would show a standard popup. But we modified the background CSS and image elements. The CSS was a clear transparent background, and the images were all blank.
Users in the control group were shown blank ads. This workaround solution enabled us to conduct A/B tests and to track the result within Google Analytics.
Using a 20/20/20/20/20 split, we ran four ad variants and one control.
Interestingly, we did not see significant differences between ads. Comparing segments 1:1, they all performed in line with the control or outperformed the control.
Notably, this test involved the entire population. Each 20% segment was large enough that we were 99% certain that the results were valid.
This particular part of the cart is a high-intent part of a user's purchasing cycle. Thus, the users exposed to the ad were at the end of their purchasing cycle.
| Pop-Ups | No Pop-Ups | % uplift | |
| Transactions | 300 | 43 | |
| Users | 2644 | 571 | |
| Sessions | 2884 | 575 | |
| Pageviews | 54,322 | 8,358 | |
| Conversion Rate | 10.40% | 7.48% | +39% |
| Avg. Order Value | $375.78 | $316.51 | +18.7% |
| Value | $112,733.92 | $13,609.98 |
By taking both the conversion rate uplift of 18.7% and the Avg. Order value uplift of 39%, our total revenue uplift was 64.99%.
Accounting for exposure differences, the control would have generated $68,366.16 instead of the $112,733.92 it ended up generating.
(2884+575) * (0.0748) = 258.73 transactions
258.73 conversions*$316.51 = $81,891.64
259 transactions - 43 transactions = 216 transactions*$316.51 = $68,366.16
If we had used no popups, we estimate that the segment exposed to the popups would have generated $68,366.16 instead of $112,733.92, a difference of $44,367.76, a 64.89% difference in revenue.
It's important to note that not all tests resulted in revenue uplift. Previous iterations were flat or even negative depending on targeting the ad copy.
With this test, we found that exit intent popups can work, especially if they are highly targeted to the right audience at precisely the right time.
Amazingly, using simple and cost-effective tools, Optinmonster is only $137 a month, and we were able to increase new customer revenue by 69%. That is an astonishing result.
If you had told me a couple of years ago that I could type "a cyberpunk hedgehog making a latte" and get a photorealistic 4K video back in seconds, I would have laughed. But here we are in 2026, and AI video generation isn't just a novelty anymore, it's a massive part of my daily workflow.
This guide leverages my experience to break down how to write, structure, and publish a document that earns trust rather than just demanding attention.
Depending on who you ask, there are anywhere from five to twenty "essential" rules out there. But in my experience, there are really only a dozen “laws” of visual design that matter across every medium. Here’s a guide I’ve created with the elements I find to be the most important, no matter your platform.
The results are surprising, Quality Score and Cost Per Click are more clearly correlated when you look at Exact Match keywords.
I love WordPress for its customizations. Styling code snippets enhances user perceptions. Copy and paste the code below to style your WordPress code blocks.
The code snippet plugin changed my life. I no longer edit the raw PHP files to add simple functions like titles to users' profiles.
If you had told me a couple of years ago that I could type "a cyberpunk hedgehog making a latte" and get a photorealistic 4K video back in seconds, I would have laughed. But here we are in 2026, and AI video generation isn't just a novelty anymore, it's a massive part of my daily workflow.
This guide leverages my experience to break down how to write, structure, and publish a document that earns trust rather than just demanding attention.
Depending on who you ask, there are anywhere from five to twenty "essential" rules out there. But in my experience, there are really only a dozen “laws” of visual design that matter across every medium. Here’s a guide I’ve created with the elements I find to be the most important, no matter your platform.
The results are surprising, Quality Score and Cost Per Click are more clearly correlated when you look at Exact Match keywords.
I love WordPress for its customizations. Styling code snippets enhances user perceptions. Copy and paste the code below to style your WordPress code blocks.
The code snippet plugin changed my life. I no longer edit the raw PHP files to add simple functions like titles to users' profiles.
