Ever wondered about the impact your Web site display has on your sales? Sure you have.
If you’re a small business e-commerce operation, your main page is very much like a display window in a department store. The catalog pages where your merchandise is displayed are akin to the shopping counters.
I’ve often wondered why eBay and Amazon have front pages that are so cluttered. They are not to my taste, but clearly you can’t argue with success. Still, I want my orchid company Web site to be more like Bergdorf’s than Wal-Mart — for purely aesthetic reasons. Of course, sales are what ultimately count, and I could be very, very wrong.
That’s why I was interested to discover an ongoing study in this area — a real-world test of various designs and merchandise presentations and their impact on online shoppers.
It’s called the Internet Consumer Purchasing Behavior Study and it’s being run by the Haas School of Business at UC Berkeley in conjunction with e-commerce services provider SysIQ and the Direct Marketing Association.
The purpose of the study: “To understand and to measure the effects on Internet consumer purchase behavior of various attributes of a Web store’s look and feel, site navigation, site features, product presentation, composition of featured product sets, effects of coupons and price promotions, as well as to measure the price elasticity of demand for various product categories.”
Kind of gets at the heart of e-business doesn’t it?
Roman Martynenko, CFO and a founder of SysIQ, told me that this experiment originated from conversations with Miguel Villas Boas, Professor of Business Administration at the Haas School.
“I was nurturing the idea for several years — ever since I realized that ‘cell testing’ (the equivalent of A/B splits) is not commonly done by small- to medium-size retailers and catalogers,” Martynenko said.
“This project was a natural outcome given Miguel’s academic interest in Internet consumer behavior and SysIQ’s expertise in hosting e-commerce solutions for catalogers and retailers.”
The experimental site, running on the SysIQ platform, sells office products (the fulfillment company is United Stationers). Customers are driven by search engine ad-word campaigns to the site, and they can actually buy products. However, some consumers get one look while others get a different presentation of the merchandise.
The study is initially aimed at three main things:
- Determining how the number of products presented simultaneously affects the likelihood of a purchase.
- Determining the ideal catalog structure.
- Determining whether, when consumers have a choice of three levels of quality, they tend to purchase the medium level more often than when the medium quality is offered by itself.
The analysis will be based on the actual consumer behavior displayed by real consumers making real purchases and the differences in such behavior, if statistically significant, observed on each version of the site.
And here’s what’s really cool: you can submit ideas for further study.
I know I’d sure like to have some answers and guidance on a lot of these issues for my online business. The study began last November and is ongoing. I’m not going to tell you how to get to the Web site because that could distort the experiment’s results.
I asked Martynenko how many different site versions are being presented to consumers, and he said it varies from experiment to experiment, but it may go as high as 20 different versions at a given time.
Martynenko said that the results would be made available via academic papers published by study participants in marketing research magazines, and less academic and more practical analysis/conclusions/implications will be published on the SysIQ site.
No preliminary conclusions have been drawn, he said, adding that he expects some of the results may well be incorporated into the SysIQ e-commerce platform itself.
An overview of the first three studies can be found here.
What’s nice, I think, is the utter neutrality of this study.
“Today, retailers and catalogers who study the Web site behavior of their customers and visitors are limited by the incentive to sell goods,” Martynenko said. “The unbiased analysis from this study will be extremely useful to an online retailer or cataloger who might not be willing to modify their site or test a certain attribute for fear of losing sales. All behavior captured is 100 percent anonymous and is used strictly for academic purposes.”
My view is that this has some significant potential to help all e-commerce entrepreneurs increase sales, so I will keep you informed when the preliminary results are in and conclusions are drawn.
At present, the study is set to continue as long as there are new experiments in the pipeline, according to Martynenko.
Meanwhile, I’m going to continue on the side of elegance with my e-commerce site. But if the study shows clutter gets more sales, I’m going to junk it up in a heartbeat.