Hunting Unicorns

Perfection is achieved not when there is nothing more to add, but when there is nothing left to take away.

Antoine de Saint-Exupery (1900-1944)

Here’s my high level analysis of the technology landscape for retail planning systems:



I’ve seen systems that intersect with any two of those circles, but I’ve never seen one in the “sweet spot”:

  • Built for retail
  • Holistic supply chain planning capabilities
  • Commercially available with a solid track record


Built For Retail and Commercially Available, Limited Supply Chain Planning Capabilities

Systems falling into this category generally have the “look and feel” that retailers are looking for and speak the language that most retailers find natural: orders. Suggesting orders. Optimizing orders. Managing the release of orders. 

Over time, these systems have evolved to include long term demand forecasting, time-phasing out their ordering logic into the future and even connecting time-phased store order plans to distribution centres in an attempt to encroach into the “Supply Chain Planning Capabilities” circle. But the logical DNA of these systems is to work out the administrative part of the supply chain (the orders) first and understand the shipments and arrivals after the fact.

While these systems are demonstrably and significantly superior to traditional reorder point and min/max approaches when it comes to replenishment, they struggle to provide a valid simulation of reality that can be rolled up to support flow planning, capacity planning and business/financial planning, particularly in scenarios where the “steady state” is being disrupted:

  • Changes to network flowpaths, such as realigning DC outbound schedules or changing the inbound source of supply
  • Properly constraining ship dates for things like Chinese New Year and scheduled supplier shutdowns
  • Properly constraining arrival dates to account for receiving schedules at stores or DCs

These systems are generally streamlined and slick, but will struggle when the following question is posed:

How would you configure the system to accurately plan for a scenario where we are currently sourcing a bunch of items domestically, but will start sourcing those same items from overseas in 5 months?

Supply Chain Planning Capabilities and Commercially Available, Not Built for Retail

Systems falling into this category trace their lineage back to manufacturing and distribution where the discipline of supply chain planning began. Planning stock movement in a backward stepwise fashion from demand to supply (i.e. demand triggers arrivals which trigger shipments which trigger orders for every item at every location) is built right into their DNA.

Over time, these systems have evolved to be able to process the gargantuan data volumes common in retail, but only through brute force and by the grace of Moore’s Law. And bolt-ons have been developed to plan for things like retail promotions and intermittent demand streams in an attempt to encroach on the “Built for Retail” circle.

While these systems excel at being able to holistically plan stock movement from source of supply to source of consumption, it only comes with unnecessary complexity. It’s not easy to genetically modify a system that was built for manufacturing and distribution into a retail solution. These systems are designed to follow the core principles of planning, but will struggle when posed with the following question:

How would a planner update their forecasts and safety stocks for 20 items across 500 locations, roll up the results and then make a few tweaks – all before 10am (on the same day)?

Built for Retail with Supply Chain Planning Capabilities, Not Commercially Available

Systems falling into this category have successfully translated the time-tested planning capabilities originated in manufacturing and distribution to specifically tackle the retail planning problem in a way that’s simple, intuitive and fast.

The biggest problem these systems face is the huge barrier to entry into the market. In spite of their shortcomings, the types of systems discussed previously have developed a track record for delivering significant benefits to their retail customer base – suboptimal planning is better than no planning at all.

These systems have everything retailers need (from a stock flow planning standpoint) and nothing they don’t. But in my experience, retailers aren’t generally known for their willingness to gamble on something new and unproven at scale. They will struggle when posed with the following question:

Tell me about your last 5 full scale implementations at a retailer our size with similar planning challenges?

If you’re a software provider (or a user of said provider) who thinks you’ve hit the trifecta, then I guess I’m implying that you don’t exist. Even though I have never heard of you, I would be thrilled to get to know you.

I’m skeptical, but I’m eager to be won over.

You know the types of questions I’ll be asking.