As the seasons change and the weather starts to warm up, I found myself on the hunt for a new summer outfit – ideally a lightweight blouse paired with a matching skirt. While the merchandise I’m looking to buy isn’t necessarily relevant, the situation I found myself in is.
It’s the kind of situation consumers experience waaay too frequently and, frankly, a large reason why providing a stellar customer experience while improving sell-through is such a challenge for many retailers today.
The story goes like this:
I came across THE perfect blouse/skirt on a retailer’s e-commerce page. Sizing is always a potential issue so instead of making an online order, I resort to a brick-and-mortar visit.
As I make my way to this retailer’s store, ready and willing to spend, I ultimately find:
- The blouse isn’t available at that particular location
- The skirt isn’t available in the size or color I wanted
- The sales associate was super unhelpful (irrelevant to the point I’m trying to make – but argh—still worth noting)
Okay, fine. It happens. How was the retailer to know I was going to waltz into their Newbury location looking for product A and product B?
The thing is, businesses “doing retail right” do know – and plan their assortment accordingly. Successful retailers know when it comes to finalizing their assortment plan for the upcoming season, they must rely on truly localized demand to ensure they stock the right product at the right location (and in this particular instance, in the right color and size).
The Challenges of Traditional Retail Clustering
Let’s take a few steps back from this scenario, all the way back to when retail planners, buyers, and allocators make the big merchandising decisions on what product to buy and how much of it to buy.
How do they determine what to stock in each store?
Traditionally, merchants rely on a clustering process – which, in the simplest terms – is the idea of grouping stores together to make purchasing inventory easier. The goal is to (hopefully) maximize profits by grouping stores with similar characteristics together to make sure they’re putting the right type of product and the right amount of product in each location.
“Clustering, when executed correctly, can help accurately plan everything from products and pricing to staffing and store design, without the added complexity of planning at the individual store level.” – Robin Report
Sounds pretty straight-forward, no?
There are a ton of different ways retailers do this, using all sorts of information—such as store size or sales volume—to simplify inventory decisions and categorize stores in a way that’s based on each store's ability to sell different types of product.
While the concept of clustering is completely on point, the methods retailers traditionally use to drive their clusters are outdated.
Let’s walk through some of these clustering scenarios, along with some of the all too common challenges presented by each method:
Clustering by Store Size
Many retailers often cluster by store size. The logic behind this method is pretty simple: a smaller store can’t hold the same volume of merchandise as a larger store.
Clustering by store size has nothing to do with consumer demand. It’s more focused on a retailer’s own space constraints. Every retailer has to account for their own unique set of constraints to some extent. However, as a stand-alone method for clustering, relying on store size isn’t enough to maximize sales and improve performance.
Clustering by Sales Volume
Sales volume is probably the most common way retailers cluster stores. For example, merchants will group stores with the highest sales volume in cluster A and those with the lowest sales volume in cluster B.
Merchants will then plan their assortment and buy quantities for products around each cluster’s previous performance, ultimately placing more merchandise at higher sales volume stores (i.e., cluster A) and less merchandise at lower sales volume stores (i.e., cluster B).
The problem with this approach is merchants make a huge assumption about product sell-through across a set of stores based purely on quantities and historical performance. The assumption is that stores that sold the most merchandise last season should get the most merchandise.
This approach is flawed because you’re not really accounting for true consumer demand, which at the most basic level is rooted in customer preference.
Understanding Customer Preference to Improve Clusters
What if you could impact your clusters based on a solid understanding of how your customers choose?
Going back to the retail experience I shared earlier:
What if the blue skirt I was looking for was available in black? Would I have made that purchase instead? What if another choice was introduced into the mix?
The entire experience could have ended differently. It could have ended with a purchase and a happy customer (along with a different blog topic for this week).
But it didn’t.
This retailer (like many) doesn’t have the insight into what products are actually driving sales and which customers prefer certain product attributes.
The clusters retailers traditionally rely on to make their assortment planning and buying decisions are based on external factors that don’t reflect what the customer wants.
Relying on things like sales volume or store size to make merchandising decisions will result in a less than optimal assortment and inaccurate quantities across stores.
Successful retailers are leveraging advanced analytics tools, like Celect, to determine the optimal mix of store assortments, across all locations by creating data-driven store clusters where customers make similar choices and stores have similar demand patterns.
At the end of the day, it's really about understanding what the customer wants. As a retailer, if you can do that, you'll be as good as gold.