An in depth look at collection merchandising strategies and insights that you won't find anywhere else on the web
Quick Jump to:
Without fail, nearly 80% of the visitors to any given ecommerce site will, at some point during their session, find themselves on a collection page. These visits to collection pages account for the majority of product views generated by a store, and, as a result, *also drive the majority of sales.* Intuitively, this makes sense. What’s the first thing you do when you visit a new store? Most of the time you click on a link in the top nav, which takes you directly to a collection page. Sure, you’ll occasionally search (roughly 10-15% of site traffic) or click directly into a product from the homepage (hero images tend to receive 10-20% of clicks), but most browsing starts in the nav because it helps you narrow down what you’re looking for. The narrowing down aspect is important for conversion as well - it shows a certain level of intent by the (potential) customer as they move down the purchase funnel.
However, collection pages are rarely optimized for conversion.
The distribution of collection views on ecommerce stores tends to follow the Pareto Principle where 20% of the collections generate 80% of all the collection page views on a site.
Practically, this means that you can get most of the benefit of optimizing your collection pages by focusing on just a handful of pages; however, figuring out which pages get the most views can be complicated because accessing collection view data on Shopify is hard.
For most stores, the top collection pages are (1) sales (2) best sellers (3) new arrivals (4) a speciality product for that particular store.
After spending half a decade+ thinking about merchandising challenges, this is still one of the facts that I find most surprising about ecommerce. While you’ll see tons of innovation and conversion optimization efforts in homepage design, email marketing, SEO, SEM, and related product sections, relatively little happens on collection pages (which are where shopping happens). Most sites display products in rows of 3-4 (on desktop), ordered by newest (or manually sorted), and they’ll occasionally use some form of faceted navigation. But why don’t brands spend more time refining and measuring the performance of these pages?
Simply put:
Through our work on Entaice, we’ve gotten to meet some of the best merchandising teams on the planet. They want to make their store’s shopping experience great for customers. But they’re put in an impossible position: managing conflicting priorities of generating revenue, selling through specific high inventory items, supporting marketing campaigns, and telling coherent customer stories. All without the data to tell them if what they’re trying is working. And when they can access the data, acting on it before trends change is almost impossible.
Google’s been around for more than 20 years, but search for a few hours and you’ll discover that there aren’t many good solutions for measuring collection pages performance. Which is crazy because most shopping happens on collection pages.
Think about that. Most of the ecommerce traffic on the entire planet shops on collection pages, but no one has a good metric to measure the success of the page or a good definition for what the page is supposed to achieve.
Across our conversations with merchandisers, most say collection pages exist for one reason: to help visitors discover products they might ultimately want to buy. With that purpose in mind, we can start thinking of a user’s visit to a collection page as part of a multi-stage funnel that’s measurable.
And, once we have that funnel, we can create rate metrics that will make it easy to compare and benchmark performance across our collection pages and evaluate their success:
The first time you’re thinking about building a merchandising strategy for your store, it seems relatively straight forward. You should build an aesthetically-pleasing assortment that shows your customers what they want to buy and gets those customers to convert. But the specifics can be hard to pin down:
With so many variables to consider, you’ll inevitably find there’s a tradeoff between flexibility and performance. You can have the most flexible merchandising tool in the world, but if you don’t have a set of go-to merchandising rules, you’ll never be able to generate learnings that lead to better performance.
We’ve already written about what it means for a collection page to perform well, and now we want to share some of the most useful merchandising rules we’ve seen and some of the places we think it makes sense to incorporate added flexibility.
Useful Merchandising Rules:
Places where flexibility makes sense:
If you look at the most of the ecommerce merchandising tools on the market today, you’ll notice that many of them offer “rule-based” merchandising as the primary method for generating algorithmic collection page rankings (https://www.fastsimon.com/technology/retail-merchandising/). And, at first glance, this seems like a great feature to have. You can ensure products with certain attribute show up in a specific sequence or promote certain brands / categories or products over others. However, this approach to sorting ignores the fact that most product views go to your first page of results, and, as a result, it can have a large, negative impact on revenue and sales.
For example, say you had a “what’s new” collection that contained both shirts and shorts for summer and you used a rule-based merchandising system to make sure shirts appeared above shorts. Every single shirt in your collection (including your worst performing SKUs) would show up above every short - even if 40-50% of your best-selling products were shorts. Depending on how many shirt styles you have, your shoppers may not discover that you even off shorts because they’re so far down the collection page.
While the are some rules that make sense (pushing out of stock products to the end of your collection), most suffer from the same flaw: they’re not able to blend the rule-based attributes with performance-based data to refine the product assortment.
The best solution we’ve found is creating “product groups” within a collection where 2-4 products get grouped together based on an attribute and then sorted by performance. So, rather than placing all shorts below all shirts, you would see 2 shirts, then 2 shorts, then 2 shirts again. That way you’re able to achieve much of the visual consistency you get out of a pure rule-based approach while making sure your high-performing styles get the attention they deserve.