Multivariate Testing

Rethinking Category Page eCommerce Presentations

As eCommerce platforms have evolved over the last 10 years, one thing has stayed fairly constant – the presentation grid used within category and sub-category pages. As a definition for “grid based”, I am referring to the type of product presentation that shows numerous products for a given category, provided in a table format.

These tables vary by retailer in terms of default numbers. Some retailers may have 4 products across and 6 rows vertically before a user has the opportunity to scroll to the next page. Other retailers might have a much more deeper product showing on the page – for example the retailer may present 10 rows of products.

Interestingly enough, in all of my years of working in this sector, I have never seen available research relating to tests within this site aspect. Many of the leading platform providers offer standard default numbers with their category grids, and their clients are likely not pressing hard enough to alter the status quo within their systems.

Limiting cognitive processing within eCommerce

Site usability often relates to human cognitive processing and how this guides user behavior. This question of “what type of merchandising grid size works best” is in direct correlation to this theory.

What we want to know is are users scrolling down the page to fully absorb the offerings for which the retailer is presenting and at what number do they most likely take “browsing” actions.

At Trinity we have recently conducted site tests that changed the number of products presented within a key site aspect in the transactional funnel. Within this test we found that a substantial increase (20%+) in conversion happened when we reduced the number of offerings that were being presented to the user.

What we did within the test is reduce the cognitive overload within the page and assist the user in digesting information. Ultimately this increased the ratio of users who took the desired action and created sales growth.

So where do you begin if you are looking to drive optimization within this site element? First, conduct some user tests that are highly focused on tasks that are realted to finding products on category and sub-category pages.

Observe how users progress through your products. Are they scrolling below the fold? Are they progressing forward to the next page of products? Understand what percentage of your test executed these two site actions.

Ask verbal questions to the respondents after and during the test.   Ask questions that focus on the location of products and how easy it was to identify items during the test scenarios. Often the verbal responses that you receive can uncover unthought “nuggets” that are correlated to cognitive overload.

You see your website daily and these respondents may be seeing it for the first time. This new set of eyes is invaluable in understanding how new visitors interact with site elements.

Now you have some primary research relating to your site which is going to help your decision making. The next step is to evaluate your current abilities within your eCommerce system and if you have the ability to change the default grid attributes (the number of rows and columns of products).

Executing this type of experiment within a split test could be difficult depending on the flexibility of your code base.  A more realistic approach would be to try an alternative layout on a week per week basis to assess the difference. When considering the KPI’s that are going to be used to identify success, consider the following:

Number of carts started
Number of product page views
Progression ratio of category to sub-category pages
Average dollar value per visit

These four metrics can provide the intelligence and data to see if an alternative default layout for your category page works better for your customers and makes their site experience easier.

Try testing your default merchandising grids and see how the numbers impact performance within your eCommerce store. To learn more about Trinity Insight and our eCommerce optimization services, please contact us online.

Path Analysis and Dynamic Funnel Generation for Ecommerce

In a previous post, I discussed the new Google Analytics and how this version has placed a greater emphasis on path analysis.  For those who may be new to the topic, path analysis is the practice of reviewing website behavior relating to browsing tendencies, in an effort to better understand the effectiveness of copy, content, and information architecture.

One of the most important tools in being able to conduct effective path analysis is the ability to generate uder defined “funnels” for processes that exist on your current site.  For instance, as an eCommerce merchant, you may have the following processes that you want to track.

 Cart – Checkout Transaction process

Search Results to Order Process

Homepage to Email sign up process

Key landing page conversion funnels within PPC advertising

In these instances, when conducting path analysis, you want to create page progressions that replicate how a user will shop your store and identify the pages that are creating the most leakage.  Omniture, a web analytics software leader, provides the capaibilties to do this exercise by dragging and dropping pages within their PathFinder report module.

By creating these page progressions for path analysis, you can identify the pages within the progression that are losing high numbers of visitors.  By taking new approaches to layout, copy, creative, images, etc on these pages – you will be taking the steps necessary to be proactive in positively impacting convesrion rates through concrete web analytics data.



Behavioral Targeting and why it is eCommerce’s most important trend…

I was in heavy notebook shopping mode this afternoon. as Ive been sick of my heavy Toshiba dragging me down at every airport and tradeshow. I was in full research mode looking for the lightest laptop I could get with enough power to run my vital apps.

In doing so, I went to many web properties. I entered ” light laptop” numerous times. Now, when I went to Circuit City (directly, not through an ad), I was provided a enticing “promo” in the homepage of the primary real estate of the site for a notebook computer.

Almost automatically, I clicked the spot and was being merchanised to on the sub-category level. Data shows that once a user has progressed to this point in an eCommerce transaction, chances of conversion dramatically increase.

The question is – did Circuit City have some type of tracking software set in place and a program with other web entities so to present RELEVANT CONTENT AND PROMOTIONS to my needs at that current time? If they did, it worked and I was fully engaged in the user experience.

This is what I feel is the next big thing in eCommerce, not flash based merchandising designs which seem to be the hot thing at eCom trade shows.

My opinion, in this instance it was completely random. Circuit City did not have the processes in place to have their site present so intelligently. But you can replicate this type of targeting on our internal site and create a much more engaging experience for your visitors and customers with current technology available today.