RSR Research: The Future of Merchandise Planning – Incorporating Customer Segments
Through a special arrangement, what follows is an excerpt of a current article from Retail Paradox, RSR Research’s weekly analysis on emerging issues facing retailers, presented here for discussion.
As many of you know, I was among the first IT execs in the U.S. to successfully implement a merchandise planning system. The planning process at that time was pretty straightforward and has remained more or less the same for the past twenty-five years. But that’s changing now in some really fundamental ways.
First: the computing power we now have available to us allows us to build our plans from a SKU-level demand forecast. That’s very different from the way we used to build them and can be difficult to wrap your head around.
But the second area — and the one I want to focus on today — is the wealth of customer data we now have available to us from online, social and mobile channels.
Merchandise plans are typically viewed as a cube of three dimensions: Product, Location and Time. When the online world came along, the e-commerce channel was generally viewed as a separate location, so it got tucked in there.
All planning systems worth their salt have always allowed alternate hierarchies in each of these three dimensions so:
- Products could be rolled up either to class, department and category, or to buyer;
- Time could be rolled up to months and seasons or, in the case of retailers who also sold through catalogs, to catalog "lifecycles";
- Locations could be rolled up geographically or by regional management.
Attributes are also frequently used to group seemingly unrelated elements. In fact, in this year’s recent Merchandising Benchmark, 61 percent of Retail Winners (those who out-perform inflation in year over year comparable store sales) rated attribute-based planning systems as very important technologies.
It’s also worthy to note that 50 percent of respondents to that same survey also said they see incorporating customer data into their merchandising processes as a top-three opportunity. But if customer data is that important, are attributes enough to do the job?
That’s what really got me thinking. I’ve run the concept of "customer" as part of the planning model through my head for several years. But it’s hard enough to think of a cube, let alone a tesseract (apparently this is a four dimensional object) for merchandise planning purposes. I just couldn’t wrap my head around how it would look or even function.
We have always picked our store locations based on how well they geographically matched up to our target customer segments. Or as the mantra goes, retail is all about "location, location, location." But in an omni-channel world, don’t we have a chance to flip the thing around? Stores aren’t going anywhere, but can’t we look at customer segments across geography, not as an attribute of a location or a product. Why shouldn’t customer segments be an alternate hierarchy of location? In fact, as the world goes more and more virtual, you’d almost expect location to be the alternate hierarchy of the primary hierarchy: customer segment.
Discussion Questions
How should customer data be incorporated into merchandise planning? If it’s a separate dimension like Product, Location and Time, what does the customer segment hierarchy look like? Doesn’t there need to be reconciliation between segments and locations, and if so, isn’t it really a planning tesseract after all?
The customer is found at the location. Insight about the customer is an essential planning factor that today’s merchants must leverage, or otherwise lose to the competition. They need to take into account their stated preferences, observed channel behaviors, frequency of visit, categories purchased, average spend and ways in which they choose to respond to stimulus.
I agree it is multi-dimensional and there are new, evolving tools/predictive capabilities to help merchants manage more insight/more data to succeed. The retailer that has the capability to adjust display and depth of product by location wins.
Retail and brand merchandising and marketing should plan their efforts from the customer’s perspective. This includes all attributes about your existing customer and prospective shopper. They are digitally empowered and they are in control of who, what, where, when, and how brands of their choosing are communicating with them through channels of their choice.
Customer attributes and segmentation are critical for retailers to provide the correct product and services in the easiest manner possible. Some retailers are in fact developing and implementing this strategy. Retailers need to question the role their store infrastructure in this omnichannel environment. If the store is not tuned into their immediate local audience and environment, they will simply become irrelevant and an expensive place to store stuff.
The consumer has been the foundation for product, location, and time decisions. However, with more data available, consumer media choices, shopping styles, communication styles, and networking must be added to traditional consumer attributes.
What we need are models that seamlessly integrate real time customer data with supply chain and other data. That’s the Holy Grail, but it needs radically restructured systems and software to get there.
BTW Paula — and anyone else who is interested — in geometry, a tesseract is four dimensional analog of the cube.
Welcome to the matrix, Paula. I believe geographic location has always been a special instance of differentiation by customer segment. By banner and by region have been the only convenient summary units until pretty recently.
Now that we are awash in a fire-hose of new customer data from social media, search, and e-commerce behavior, merchandise planning at the store level seems to be imperative. Since it can be known, it must be incorporated, right?
The first hard part is profiling each store so that each maps to relevant variations in the shopper base. The second hard part is implementing consistently against the differentiated plans, while collecting store-level performance data that may be used to refine the next planning cycle.
Non-trivial stuff, to be sure. It may be overwhelming for large chains, due to its sheer intricacy and the remoteness of stores from the HQ. Smaller chains may simply lack the resources to do the data crunch. But there’s no avoiding the multi-dimensional mandate of the omni-channel world. Winners will work out the necessary practices.
By the way, “Tesseract” would be nice name for your next chain of boutiques, don’t you think?
This is such a smart approach and is indeed where merchandise (and marketing) planning SHOULD be today.
Customer data, especially past purchasing behavioral data, is the ultimate predictor of future behavior and thus a way to forecast the demand side. Whether a consumable, a staple, or a fashion item, there are factors to consider such as frequency in a given merchandise department/class/sku/etc. and the total volume in that category. Considering these data elements, on a same-customer basis, can be modeled as demand for merchandise planning.
The challenge is that most retailers are not good at using customer data to begin with. While there are some exceptions (Amazon, Best Buy, increasingly — Macy’s, J. Crew), looking at the lack of data-driven relationship marketing is testament to this.
This is a very interesting article with a lot of fuel for future seminars and workshops. I too have a long-time working relationship with the IT industry and have a good handle on the amount of and quality of information that exists today. The relatively cost-effective transportation and communication systems in place throughout Europe and North America make consumer location far less predictable than it was even 20 years ago. It is this reason that retail is finding it harder and harder to find viable locations outside of mega metropolis areas for a brick site. For that reason I think we will see smaller stores with a lot more IT catalog deliveries. The biggest stumbling block is the “where can I see it” objection which will be worked out in the near future. The issues and challenges at hand are more about the quality and use of data, as well as how to create concise meaningful reports. Therein lies where today’s executives are stumbling.
Good article. At dunnhumby, we’ve been using customer data in merchandising planning for some time. It’s not that straight-forward to be sure. For example…
1. product elasticity curves for loyal shoppers instead of all shoppers
2. how customers make category choices, e.g. do they decide first on pack size, then variety, then brand, or is it something else? How does this differ by store?
These can be done quite easily on a one-off basis, but are hard to scale across an entire enterprise.
Part of the challenge is around data and technology, but a broader part is on the organizational readiness to make customer-centric decisions at this level and be rewarded in this way.
To drive breakthrough value, the same customer data and views need to be applied consistently across all levers the business has, and with suppliers too. It is only when everyone is going in the same direction with decisions on price, assortment, promotions, communications, service etc. that the customer experience changes and they consistently choose you more often.