how to increase revenue with multiple split tests

Case Study: Increase Your Revenue With Multiple Split Tests

andrew browne Product Listing Optimization, Split Testing Best Practices 1 Comment

Sharing is caring!

Have you settled on some strategies for increasing your Amazon revenue in 2017 yet?

Since the Terms of Service changed back in October 2016, Amazon FBA and private label sellers have definitely had to adjust and move with the times. One very successful strategy to introduce is split testing, or conversion rate optimization.

Several of our most successful Splitly users have tested their listings several times, to make sure they are fully optimized. Many however, stop at one test, which means they are still leaving money on the table.

We wanted to share some results from a success story. A case study of someone who was able to increase revenue again and again by running several tests. In fact, this person is now taking an average of nearly $10k per month more than they used to.


Why you should run multiple tests?

One test is not enough to optimize your listings. This is because there are many listing features that can be tested such as price, featured images, images, title, product features and description.

In addition, each listing feature can have thousands of variations. For example, you can test several featured images to find the ones that drive more traffic to your listings. Similarly, you can test several price points to see which raise conversions. You can also test several different copy styles for your description and bullet points. The possibilities are endless.

Moreover, things change over time. You may test something on your listing which helps increase your sales metrics. However, after a few weeks, this test may no longer be relevant as the Amazon landscape may have changed.

For example, your niche may have seen an increase in competition, or perhaps your competition started to reduce their price to compete with you.

Therefore, you must test the same listing features regularly to make sure your listing is fully optimized.


Case Study: $9,491 increase in monthly revenue

One of our users was able to increase their revenue by nearly $10k a month after running multiple tests.

This seller has over 8 products and ran over 50 tests. This is because they tested several products, and also several listing features more than once.

Below we will run through just some of these tests to see how they have improved revenue over time.


Successful Test 1: Standard price test and sale price

This was a multivariate test, which means more than one variant was being tested.

One variable tested was the standard price, with an original standard price of $24.79 and the test price of $21.99.

The second variable was the sales price. The original sales price was $18.99 and the test price $19.99.

The standard price change caused a decrease in sales, and therefore was not successful.

However, the change in sales price from $18.99 to $19.99 increased sales  by the following:

  • Daily sales average increased by 17%
  • Conversions average increased by 19%
  • Daily profit average increased by 29%
  • Conversion rate increased by 18%


multivariant pricing and sales price test results


Tip: This test shows that you don’t always need to decrease your price to increase sales. In fact, sometimes increasing your price can cause a spike in sales. Test all price points before assuming anything.

Successful Test 2: Close up feature image

Another test this person ran was to change the feature image (main image) to make the product clearer. This was simply done by using a photograph that was more of a close up than the previous image.

This new feature image used more of the space Amazon allots to display more of the product. This change increased average daily sessions by 18%.

Tip: Use most of the room Amazon gives you for your featured image. The less white background space the better. Make your product images big and bold.


Successful Test 3: Change in feature image

Not satisfied with just one test on the main feature image, let’s see what happens when the image is changed again.

This time the image was still close up with the product clearly shown. However, the seller added some value by showing a bag that was included with all purchases too.

This led to a further increase in average daily sessions by 27%.

split test changing feature image


Tip: Include any value adds you give in your featured images.


Successful Test 4: Price test (again!)

This time, a different product is having a pricing tested from $24.99 to $22.47. This price change caused most daily sales metrics to increase:

  • Sales average increased by 58%
  • Conversions average increased by 64%
  • Daily profit average increased by 33%
  • Conversion rate increased by 54%


Tip: Try testing several price points including increasing your current price point and decreasing your current price point.


As seen in successful test 1, a price increase may increase your sales. However, as seen in this test, a decrease in price might also increase sales.


You can automate your pricing too…

In fact – because pricing can have such a profound effect on your sales and profits, and because this fluctuates so much, we introduced Profit Peak.

Profit peak is essentially an automated repricer for private label sellers, which will always be figuring out your best known price. This means the software can react when your competitor changes their price, demand falls or increases or seasonality affects the best known price.

Running Profit Peak on Splitly does mean that you can’t run any other tests on the same ASIN, though. So if you want to optimize your images, copy or anything else, it would be best to do this first. Then you can turn on Profit Peak for ongoing pricing optimization.


Successful Test 5: Picture orders

This was a multivariate test to try out several different picture orders. That is the order in which the carousel of images appear in the Amazon listing.

This user’s first order resulted in the following increase in daily sales metrics:

  • Sessions average increased by 26%
  • Pageviews average increased by 38%
  • Sales average increased by 15%
  • Conversions average increased by 23%
  • Daily profit average increased by 13%
  • Conversion rate increased by 1%

The second variable resulted in the following daily sales metrics change:

  • Sessions average increased by 22%
  • Pageviews average increased by 28%
  • Sales average increased by 25%
  • Conversions average increased by 42%
  • Daily profit average increased by 23%
  • Conversion rate increased by 16%

multivariant price test results in splitly

Although both variables did better than the original, the second variable proved to be the best overall. This is because the second variable increased the conversion rate by 16% and profit by 23%.


Tip: Test your image order several times. Picture order can heavily affect sales metrics, and should be tested several times until you find the optimal order.


Tip: Try putting your best pictures first. Also try putting infographics and lifestyle photos at the beginning.



What about tests that don’t improve profitability?

As we can see, this person ran many successful tests. This included even more featured image tests and price tests that were not included here.

However, a number of tests proved that the original variable was more optimized (read: more profitable) than the test variant.

These tests are just as important as tests that do increase sales metrics. Even though it can be a little disheartening when a test result shows your variants have not worked. The point is, you will now know that your original variant does work, and you can continue to test from there.

Running split test experiments like this prevents you from making a costly mistake with your Amazon listing.

Here’s an example of a test where the variant did not improve profits. This was a test on the order of the product images:

  • Sales average went down by 22%
  • Daily conversions average went down by 15%
  • Profit average went down by 22%
  • Conversion rate went down by 16%

split tests that didnt improve profitability


Now this seller knows that the original picture order is more optimal. They can now stick with the original, or test again with a different variant.



If you are looking for a way to make measurable improvements, with reliable results, then consider implementing an ongoing split testing strategy. We recently shared a strategy for the whole year in 2017 which you can find here. This should help to give you some ideas when getting started.

If you are already a Splitly user, and have only ran one or two tests so far and need our help, just let us know by emailing [email protected].


The moral of this story is: the more you test, the more likely your listings will be optimized, which means you will take more profits!



Sharing is caring!

Comments 1

Leave a Reply

Your email address will not be published. Required fields are marked *