Split testing for amazon sellers
Download as PPTX, PDF0 likes6,933 views
This document discusses the importance of split testing for improving conversion rates and sales in an e-commerce context. It outlines a methodical approach to conducting split tests on product listings, emphasizing continuous adaptation to changing market conditions and the significance of statistical analysis. The document concludes with a call to use data-driven strategies to gain a competitive edge.
Split testing for amazon sellers
- 1. THE MILLION DOLLAR CASE STUDY Session #15: How To Split Test For Improved Conversions Journey to $1MM May 31st, 2017 100% to
- 2. What is split testing? Controlled Randomized Experiment Between two or more variations... 2 Will lead to Improved Metrics (ie click through rate, conversions, profits, etc)
- 3. 3 Reasons Why Split Tests Are Invaluable Improve Conversion Rate: An increase in your conversion rate will increase your sales, and the increased sales velocity increases your rank, and gets you even more sales. All of this without needing to spend on PPC or reduce your price etc. Get More Visitors: An increase in your sessions (from CTR or keywords) has a similar benefit. Price Wisely: The ROI of having your price where it’s most effective can be enormous.
- 4. The Trick Of Split Testing: Be Methodical It’s like flipping a weighted coin, the more times you flip it, the more sure you can be which way it’s weighted. Split testing is an ongoing process. Market conditions, seasons, customer demand, competition, (and more) are constantly changing. So best change your listing with the different versions many times, as opposed to just once.
- 5. Split testing, Amazon style: Amazon only gives traffic data on a daily basis (therefore must make changes daily at 12 midnight) Sales velocity affects search rank (30 day rolling average) 5
- 6. How to do a split test manually 6 1. Decided on the different versions of your product you would like to test. We recommend keeping it simple here and just having 2 versions, an A/B test. 2. Wait until midnight of your listings Amazon Marketplace; This would be PDT for amazon.com 3. Referring to these versions as variant A, and variant B, update your listing to variant A. 4. The following midnight, update your listing to variant B. 5. Keep repeating this process, we recommend for 2 weeks, so as each variant gets one day of the week. 6. In seller central, business reports, “detailed sales and traffic by asin”, set this report’s date range to one day, corresponding to the first day of your test. 7. Record the sessions, page views, total orders, units ordered, and conversion rate. 8. Do the same for each day of your test, noting which data corresponds to which variant. 9. At this point, we can see which variant is winning, by simply adding up the data for each of them. However it’s important to calculate the statistical significance. The easiest way is to just input your data into an online calculator 10. A statistical significance of above 90% is pretty reliable. If it is less than this, we recommend continuing your test some more.
- 7. What Should You Test? 1. The best products to test are your best sellers (assuming multiple products). 1. Greatest lever for Sales and Profits growth involve a Price or Image test
- 8. What Affects Sessions On Your Product Listing? Title Main Image Price
- 9. What About Your Detail Page Affects Conversion Rate? Title All images Price Bullets Description
- 10. Testing Main Image affects both Sessions and Conversions Rate VS 10
- 11. The Split Testing Process Create Your Hypothesis 11 Identify What/How To Test Run Your Experiment Measure Results Rinse & Repeat
- 12. What’s the bottom line? 12 Use a data-driven approach to outperform your competition.
- 13. Use code ‘Million’ for 25% off your first month on A Small Gift
Editor's Notes
- #11: Example of something effecting conversion rate and CTR