We all know that the faster an eCommerce site is, the higher its conversion rate is likely to be. While that’s true in a general sense, in specific cases a law of diminishing returns operates. In an interesting article on the Yottaa blog, Alex Pinto considers the case of Etsy, the popular marketplace for handcrafted and vintage products.
Etsy was looking for techniques that would increase their conversion rate: they tried infinite scrolling operating under the assumption that putting more results in front of users would increase conversion rates and they tried to make sure that their search functionality was as fast as possible. Surprisingly, neither “optimization” produced any increase in conversions as measured by A/B testing against a control group.
In a further experiment, the Etsy team imposed an artificial slowdown on their search results of 200 milliseconds. Conventional wisdom would lead us to expect lower conversion rates, but, in fact, there was no change at all.
Lessons To Be Learned
Site performance is certainly an important part of conversion rate optimization. Slower sites do generally perform worse than faster ones, but at a certain point, a website is fast enough and any extra improvements will not produce desirable results.
Premature optimization is a term from the development world that refers to the process of making code faster when there’s no evidence that it’s causing any problems in its current state. For example, a developer might spend a disproportionate amount of time optimizing a function when that function is rarely used and has a trivial impact on the overall performance of the application for most users. This matters because developers have limited time and resources. To make the most of their resources they should focus on tasks that have a real-world benefit, rather than just a theoretical benefit.
The position of many eCommerce retailers is analogous. If you’ve already got great hosting that performs well with your eCommerce app, are using a content distribution network to reduce latency, and have optimized your site in accordance with best practices, the theoretical benefit of reducing load times by a few milliseconds is unlikely to be a good use of resources and time that could be better spent devising and testing on-page conversion rate optimizations.
As the Yottaa article points out, implementing standard assumptions about conversion rate optimization without extensive testing can lead to improperly aligned priorities. What’s important is to make changes that have a real and significant impact on conversion rates, and the only way to figure out what will make a real difference is through extensive A/B testing. Metrics like site performance are a proxy measure of what’s really important, which is user satisfaction and conversion rates.
If you’re pouring resources into an ongoing effort to shave ever more milliseconds off the time your eCommerce site’s pages take to load, then perhaps it’s time to take a step back and ask yourself if it is really making any difference to your conversion rates and if your time might be better spent elsewhere.Posted in: Magento