Oliver Palmer, Conversion Rate Optimization Consultant of oliverpalmer.com, joins the eCom Ops Podcast. He will talk about eCommerce product optimization, A/B testing strategies, and user research for the best eCom results.
Key Facts
- Oliver’s LinkedIn
- Oliver is an experienced entrepreneur and optimization specialist
- oliverpalmer’s website
- oliverpalmer.com is a consultancy that helps enterprises become more adept at running large-scale business experiments . Clients include Kmart Australia, JB HiFi, Nissan, and Australian Super
- Oliver holds a B.A. (Media and Communication) in Communication and Cultural Studies from the University of Melbourne
- Oliver is based in Melbourne and works with clients worldwide to grow conversion rates and improve the ROI of optimization. He ran his first A/B test in 2008 while contracting to a London law firm. Over the next few years, he worked in various eCommerce roles, from doing site merchandising at Amazon to managing the online arm of a cult magazine retailer in Melbourne
Key Takeaways
- [ 00:06:27 ] To get valid statistical inferences for A/B testing in eCommerce, you need to have a lot of traffic
- [ 00:07:02 ] To assess an A/B test for statistical significance and validity is to look at the baseline conversion rate. The higher your conversion rate, the lesser your sample size
- [ 00:10:30 ] For small and medium-sized companies with lesser traffic and revenue per annum, A/B testing is probably not necessary; user research will suffice to get the desired results
- [ 00:18:36 ] The major mistake optimizers make is not having a good hypothesis, for example, enlarging the hero banner’s size to increase the website traffic. After close analysis, they find out their hypothesis is wrong
- [ 00:21:02 ] Even if you are doing A/B testing, you should also carry out user testing alongside
- [ 00:21:50 ] Utilize A/B testing, not for discovery but for validation
- [ 00:21:55 ] Whether or not you are running experiments, make sure you are talking to customers and doing qualitative research
- [ 00:22:30 ] Alternative to talking directly to the customer is to talk to your customer service and sales team to find out your customers’ pain points
- [ 00:22:36 ] Whatever you do, get close to your customers before running A/B testing. This will provide an extra level of validation and ensure you have a high level of success
- [ 00:23:20 ] Concerning green/red button tests, the only time this can be successful is if there’s a color contrast and the copy on the buttons is compelling
Oliver’s No.1 eCom Operations hack
“For Small and medium-size companies with lesser traffic and revenue per annum, user research will suffice over an A/B test to get the desired results”.
Quote
