"I would certainly recommend Rob to anyone" - Touker Sulyeman Brands

How We Made £26,468 p/a With A Single amazon Experiment – Step By Step

MrPrime.com is a 20+ person agency, run by Robert Prime who has 25 years ecommerce and amazon experience.

The Goal

To use amazon experiments to test the title length of a client's SKU to see whether a longer more keyword driven title was better in terms of ROI or a shorter, more customer friendly title.

What We did

We just created a client £26,468 per year extra on a single Amazon SKU for 30 minutes work! 👇

MrPrime.com runs experiments on all of our client accounts. If your agency is not doing that, or you are not doing that, you should be.

Below is what we did, how we did it and the result which made our client £26,468 on one colour of a product they sell.

The client is in a competitive space. We wanted to test longer, more keyword driven titles versus smaller titles for their main selling product and to measure the results.

How We Did It

  1. Went to brands > manage experiments in seller central
  2. We added the ASIN we wanted to test
  3. We added the title we wanted to test versus the current title
  4. We added the hypothesis (so we could easily go back to it, simply shorter less keyword driven title, more benefit driven).
  5. We set "no end date" as the length of the experiment
  6. We submitted it!

What The Result Was

14 days later we had some solid evidence from amazon, which showed:

  • Title B, our shorter title, had a 96% chance of being better
  • The conversion rate from search (not clicks) went from 0.96% to 1.36%
  • Units sold in the period went from 65 to 93
  • Sales went from £702 to £1,021
  • Each product had over 6500 impressions as a sample size
  • As mentioned there was a 96% statistical confidence that version B was better. The average outcome would be that version B increased sales by £26,468 per year. Worst case it went down by £1,030 and best case sales went up by £53,986 per year for this SKU!

The best part is that this was the weakest colour of an identical product with 2 other colours. By changing all the colours to our new title, we potentially made an extra £100k per year.

Amazon experiments run by showing version A or version B half the time. Not one day version A and the next day version B or the results could be skewed by seasonality, day of week. They literally switch them constantly to get statistical accuracy.

What We Learnt

  • In this case shorter and more customer friendly titles outperformed longer, SEO driven titles.
  • Even an increase in conversion from search of only 0.4% increased sales by £26,478 on average. Showing relatively small increases make a substantial difference to the bottom line.

There is no cost to run experiments and the data you get back is amazing.


You can run experiments on:

  • Main image
  • Title
  • Bullet points
  • A+ content

To run experiments you need to have your brand registered on amazon.

You can run experiments on ASINs/SKUs that have enough traffic to give a statistical outcome.

Extra Insight From MrPrime

I find that the title experiment and main image experiments are far more impactful.

By measuring changes and learning data, you are able to apply the learnings across your other products on amazon and even off Amazon.

It is the easiest way to get growth I have seen, even on very experienced and high selling accounts.


Robert Prime

Owner at MrPrime.com Amazon Ranking, Growth and Support Services. Co-owner of LoveReading network the U.K.'s largest book recommendation site. 22 years in eCommerce business. Forbes Business Council member.

Make your next season the best yet

Talk with Robert Prime and find out how we can help you

Or send us a message

    *By submitting your email address, you are agreeing to all conditions of our Privacy Policy.

    Ads certified.

    Our Amazon Ads team has demonstrated their proficiency in Amazon’s sponsored ads with an emphasis on retail readiness, sponsored ad strategy and campaign optimisation. Our marketing professionals are certified to use Amazon Ads sponsored ads via the advertising console and Seller Central.

    ID: 3836032