For decades retailers have been managing their merchandise assortment, planning, and allocations based on prior sales numbers, emerging trends, store profiling, and overall demand. With continued access to richer data, retailers have never been more informed about their business.
Analyzing baskets and prior purchase behaviors is going to a new level. The amount of available proprietary, inherent, and market/demographic data including weather patterns, health/fitness activity, cultural trends, and geopolitical and economic indices can all be used for more accurate planning.
But more data doesn’t necessarily mean it’s the right data. What if retailers could more accurately pinpoint by location not only what sold, but also what is most likely to sell in the future using variables including predictive triggers? And all that before the planning process even begins.
No need for a magic eight-ball here. Reverse engineering your planning focuses on specific levers that prompt the process from the endpoint of the consumer basket all the way back to sourcing. Then layer on the capabilities of today’s artificial intelligence and you have a new model for end-to-end business planning.
Is it too good to be true? Not really. While not all retailers are ready to take advantage of this inside-out approach, many do have the ability to strategically migrate to reverse-engineered planning.
How It Works: The Six-Step Process
- A retailer establishes a key performance lever that serves as a behavioral indicator. Here is where AI/ML can be utilized early in the process to elevate and identify key strategic levers.
- The planning process starts when that lever is triggered.
- The associated values assigned to that lever feedback across functional teams provide a single real-time view of the opportunity.
- Design, merchandising, planning/allocation, finance, marketing, and store operations are notified and aligned. A single source of truth/KPIs is shared from a unified/centralized source.
- A business baseline is reestablished for proactive planning.
- The process repeats and with machine learning and AI, gets faster, smarter, and more accurate.
Why Is This Different from Legacy Planning Processes?
Reverse-engineered planning is an inside-out view. Most standard processes focus on gaining customer data, aggregating the data, usually owned by marketing/IT, and passing it along in a linear fashion based on previous behaviors or actions from the inside out. This information is shared with sales, finance, merchandising, and related teams to act on. Each function executes that plan and the business leverages a hindsight meeting each season to review and readjust for the next 12-18 months.
Typically if a retailer is selling a house brand red t-shirt and it sells well in size large short-sleeve, more of the same red t-shirts are made. If something pivots during that cycle, however, retailers may be faced with poor sell-throughs, excess inventory, and lower margins/profit. This is a reactionary approach that doesn’t make sense in today’s digital marketplace.
If that same retailer had a specific lever that was triggered with applied values predicting that certain markets/stores are likely to see a shift in color preferences, they could respond before sourcing/manufacturing and pivot with minimal to no impact on sales.
What’s a Lever?
Many industries have inherent levers built into their businesses. Retailers with insurance affiliations, for instance, know the age, patient records, coverage levels and prior procedures of their customers. If you know that certain customers with better insurance coverage exhibit higher levels of spending in general, as a retailer you could effectively re-merchandise assortments to them to skew to higher price points. By the way, this might be why Apple/Google are actively looking to gain health insights on their customers via smart device technology. Amazon, always the forerunner, is already integrating this type of customer information.
Similarly, language and ethnic demographic information by market can be used as levers. If you are a retailer with a large Hispanic clientele and rely on Spanish-speaking service and store associate teams to engage with your customers, you can correlate prior purchase color and size assortment history in these markets and apply them to other markets that correlate with cultural shifts. Meaning, emerging Hispanic markets are likely to exhibit future behaviors that are similar to existing Hispanic markets. Having early visibility of these market changes can allow you to proactively plan for forthcoming opportunities.
The auto industry is also poised to take advantage of reverse planning. Dealers have lots of information on the marketplace and demographics of their buyers that can be correlated to data on the value of their cars. Do customers have child safety options? What colors were sold to whom, how many miles are they tracking, did they finance or buy outright, and what type of repairs have they performed? How can these be used as levers?
AI as a tool can establish a series of values associated with a mileage threshold by market/dealer, consumer segment and season. For example, the dealer can adjust deliveries if a lever indicates the closing of a key employer in the area which may impact salaries and discretionary spending by demographic level. They can connect birth rates by zip code with size of vehicle and vehicle features.
But What About Privacy?
By now you may be asking how privacy regulations impact reverse planning. Clearly, retailers can no longer rely on third-party and cookie-collected data attributes/behaviors. This is where proprietary loyalty programs, store credit cards, affiliate partnerships and permission initiatives come into play. Even marketing response, channel preference, platform engagement, payment methods (pay-over-time, charge, cash), first item purchase and service participation (design services, custom tailoring, other) can be used as levers.
The process would be to identify a lever, assign values based on aggregated data patterns supported by ML/AI, look for correlations, triangulate patterns, develop a profile, build a plan, create scenario plans and keep the process dynamic and fluid. This is the opposite of qualitative intuitive levers (we sold red t-shirts last year, so let’s buy more this year).
A single trigger may be oversimplifying the concept of reverse planning, but it does help to demonstrate how specific information at a granular level can be utilized to build a cooperative and unified framework for more accurately predicting and anticipating scenarios by market, customer, and season. Once you operationalize the process and apply ML/AI, a retailer can input, aggregate, dissect and identify levers and related actions with momentum.
Then What?
Analyzing baskets and prior purchase behaviors is going to a new level. The amount of available proprietary, inherent, and market/demographic data including weather patterns, health/fitness activity, cultural trends, and geopolitical and economic indices can all be used for more accurate planning. Too complex for Excel spreadsheets and CRM platforms, new developments in AI/ML are absorbing huge amounts of information to be processed making it available real-time to multiple operating functions.
The key is to systematically link relevant data with functional areas to create a holistic vision and plan. This level of holistic planning now enables the retailer to drive, measure and execute their strategy across the board from assortment to store/staff planning. The holy grail of knowing the right products to sell to the right people in the right place at the right time is now within a retailer’s reach.