Data. Data. Data. It seems all that I read about these days is DATA. And how data, plus all things tech and digital are going to save retail. I do not dispute that. Retail is experiencing a massive evolutionary moment, accelerated by the pandemic. Retailers mired in prior century thinking and modeling are having to reinvent their game as quickly as possible. So absolutely, I embrace maximizing how we gather and use data in all its forms.
The context of my work is apparel retailing, which means we are talking about fashion and seasonality. And that immediately translates into constant change. Either the fashion is changing, or the season is changing, or both are changing. So, with that much change happening on a constant basis, how valid is last week’s or last month’s or last season’s data in predicting the future? How far into the future can I predict using fashion or season sensitive information? Can my supply chain handle in-season replenishment of fashion best sellers? Turns out…data has a shelf life…!!! And depending on the supply chain’s capacity to respond in-season, the data is valid for a tight window, or it becomes outdated very quickly. Or…it might be valid as a baseline for next year’s comparable season…depending on the RISK level of the product. RISK level meaning level of Knowns versus Unknowns.
I define 4 levels of RISK as follows:
Basics = Lowest level of risk. Seasonless. 12-month shelf life. Known, established styles, fabrics, colors. Highly predictable.
Seasonal Key Items = Moderate risk. Seasonal. 6-month shelf life. Known and established styles and fabrics with seasonal and trend color. Low/moderate levels of pattern and trim.
Trend = Upper moderate to high risk. Introducing novelty fabrics and higher level of color. Fashion elements become more dominant. Larger and more intricate embroidery, applique, and printing.
Novelty/Fashion = Highest level of risk. Highest level of novelty. Highest level of trim and color. Fashion elements are very dominant. Event specific. Wow factor. Highlight item. Meant to amuse and evoke emotion.
If I lay all that thinking into a simple quadrant it looks like:
If I then lay that thinking into a simple linear progression of increasing risk, complexity, fashion, and manufacturing inefficiency it looks like:
And finally, if I drop in real-life product, it looks like:
These items happen to be Polo Ralph Lauren styles. Note the retails that accompany the increasing levels of fashion and novelty. And note the level of predictability you could apply to sales of the BB item year after year…after year, compared to how you would predict the performance of the FF item, regardless of how well or how poorly last year’s FF item did. Each year’s data on the BB item becomes an update and refinement on last year’s data and can be used in projecting next years sales. Each year’s data on the FF item is stand alone and will be of little use in projecting the performance of next year’s styles. Data has a shelf life depending on the risk level involved.
I can summarize the role Data plays at the various RISK levels in the graphic below:
In the green bands of the business, BB and BF, we have good to very good data and good predictability. In the orange and red bands of the business, FB and FF, the historical data is less reliable in forecasting the length and magnitudes of trends, and so we have diminishing levels of predictability. So, when we go to edit and assort the entire portfolio of product, what is the best balance to create? Lean into low risk, higher predictability and profitability, or lean into higher risk, lower predictability, and risk a higher degree of margin erosion…??? If the business leans too heavily into low-risk product, the story is boring and undifferentiated. That doesn’t bode well for brand longevity. If the business leans too heavily into the more high-risk product, we do indeed have a differentiated story to tell, but that doesn’t bode well for profitability.
So obviously we are left with a series of decisions and trade-offs. How much of each bucket of RISK creates the ideal assortment portfolio? And the answer is…it depends. It depends on the brand and its Brand Promise. Gap has a very different Brand Promise than Urban Outfitters. Nordstrom has a very different Brand Promise than JCPenney. Each brand, each retailer must make individual decisions about their optimum mix relative to their customers expectations. But every brand and retailer can adopt the logic here to use as a template for their merchandising strategies. And every brand and retailer can explore new thinking and processes that might be able to inject more data into the orange and red bands of the business (yellow circle in the graphic above), and take some of the risk out of the higher levels of fashion and novelty. Because that’s where great brand storytelling and differentiation live. And every brand and retailer can do this through testing and compressing their Time/Action calendar. There are digital solutions in the market now that can greatly reduce the design portion of the process, and then the Supply Chain and Logistics teams have the challenge of working with the factories and delivery networks to get product manufactured more quickly and then move from the factory to DC more quickly.
To attack and manage this endeavor, I use a straightforward progression of thinking and logic, graphic below, and the first 3 steps are all about Data. Learn is Discovery, Due Diligence and collecting Data. It is starting to ask the question “WHY?” and look for root causes. Know is turning the Data into Knowledge, connecting the dots, and understanding the underlying “WHY?” Plan is turning that Knowledge into a forward sequence of decisions, actions, and objectives. The Plan understands the probabilities of the bell curve, the possibilities of the unknowns, and constraints of time and logistics. The Plan must embrace RISK (noted in red), for there are very few certainties.
In this train of thinking, Plan precedes Solve, which is simply short for Solve the Problem. Or in the case of apparel, Design the next collection. Well surely you can’t Plan the details of how to Design the next collection…? No, of course not. The plan creates the framework for the designer to fill in the details. The data is suggesting what items classifications are growing and what are shrinking. At the micro level, Learn and Know are providing details on bests selling fabrics and colors and trim details. At the macro level, broad market research for emerging trends is an ongoing, never ending process.
When I read about data, it is often about how data can enhance some aspect of the retail Process. And I say Process as opposed to Product. Those are the two big buckets I tend to look at…Product and Process. Product is what goes on the shelf, or website, and what the customer is ultimately looking to purchase. Process is everything that gets that Product to the shelf, and then even the Last Mile into the customer’s home. So, the whole supply chain, logistics from factory to port, in DC processing, in-store visual merchandising and presentation, order fulfillment, self-checkout, and last mile delivery are all part of the Process of retail. And I am sure my list is way too abbreviated. So given the length of this list, it is not surprising that so much would be written about how data can help make it all faster and more efficient.
But I am a Product guy. So, my question quickly becomes about how the Process can help make the Product better, more customer-driven and more profitable. And that very quickly boils down to the overall Time/Action calendar of the process. Where can time be saved? How can the calendar be compressed? Data has a shelf life! The point is not simply getting product to market faster but getting product to market in a time frame when the data that’s being reacted to is still valid. How can that speed to market be achieved? My suggestion is that the early focus on time compression be directed at the 2 higher risk product categories. There is already pretty good predictability in the BB and BF levels of the business, so compressing the calendar won’t add a lot of predictability to those layers. But if the calendar for the FB and FF levels of the business can be compressed, and if those layers of the business can be operated on more of a test-and-respond basis, then the business gains in predictability and profitability without sacrificing great storytelling and differentiation.
Here is where the Supply Chain and Logistics teams have their assignment. And I know these teams are already working hard to make the whole process faster, more efficient, and less costly anyway. I’m suggesting a parallel product-specific strategy that addresses the goal of SMART FASHION, not just fast fashion.
This thinking is not about changing the core aesthetic and attitude of any brand or retailer. It is not remotely a critique of the design or creative process. It is simply an approach that recognizes the profit opportunity by better understanding the customers appetite for the higher risk levels of the assortments that any brand or retailer offers. I arrived at this thinking by watching how retailers dealt with end of season residue inventory. By the end of any given January there seems to still be an appalling amount or prior season inventory still on the floor. The end of January is when the bulk of post-Christmas sale business is on the books. So, any prior season inventory a retailer still owns at the of January is going to be sold at a significant loss. This whole exercise is aimed at reducing the level of that unsold inventory and the composition of that inventory. If higher risk product can be bought closer to need and based on more data, I think the profit implications are huge. And the Supply Chain and Logistics teams can play a big role in contributing to the time savings needed to make this all happen.
I started by saying this was all about Data + Design. But that is now an incomplete headline. I’ll use Deliver as shorthand for the whole Supply Chain and Logistics process, so we now have nice alliterative expression for the elimination of silos! Teamwork!
DATA + DESIGN + DELIVER
