Skip to content

Analytics Meets the Cardboard Box

“Do you have a bag?“ or “Would you like a bag?” have become common parts of the brick-and-mortar retail transaction.  Reusable bags, or simply doing without, have reduced the flow of plastic and paper into recycling.  

Cardboard Box

E-commerce is a different matter.  I just unpacked a box of wine, and dealing with the protective spacers and heavy-duty box took 10 minutes and consumed most of the space in my recycling bin. 

The shipper,, is a California startup that “disintermediates” the wine business, the way Uber and Lyft disintermediated the taxi and limo businesses – by substituting analytics and AI for middlemen.  For example:

  • recommender systems boost sales
  • network flow algorithms optimize shipping routes
  • linear and integer programming algorithms optimize fulfillment center operations
  • predictive models anticipate demand, permitting optimal stocking of inventory at fulfillment centers

The (over)use of packaging materials has not, I think, received such attention.

Spot the Exception

AI-driven robots and conveyor belts are showcased in this video of an Amazon fulfillment center.  In it, you can also spot the one technology that looks much the same as it did when it was invented in 1895 – the corrugated cardboard box.  You’ll see workers folding up pre-creased boxes to ship out goods, and you can catch a glimpse of stacks of cardboard boxes of various sizes.

Anyone who shops online is familiar with the tremendous volume of cardboard boxes, and associated packing material, that just one household generates.  The production of waste material is boosted by the common practice of shipping small items in large boxes, perhaps for want of a readily-available right-sized box.  Large ecommerce merchants have algorithms to consolidate multiple items into a single shipment, but this is largely to minimize costs. Otherwise, they are not really concerned with the issue of what the consumer does with all the boxes and packing material.

A Prediction

In the near future, analytics will be applied to this “cardboard waste” problem in several ways:

  • Cardboard boxes will be manufactured closer to the point of packing and shipping, to allow closer matching of box size to product size, and reduction of cardboard box inventory.
  • Predictive models will help bridge the time gap between (a) assembling product(s) for shipping and (b) making the box (i.e. will allow “just-in-time” production of boxes).
  • Algorithms will let ecommerce customers optimize a third shipping dimension at checkout – “minimize cardboard,” in addition to time and cost. 

Just-in-time production of right-sized packaging is already here, for shippers operating at smaller scale than Amazon.  Packsize offers a range of cardboard box machines that shippers can use on premises and integrate into the shipping process.  Their top-of-the-line X7 determines the size of the item to be shipped, cuts a box to those dimensions, places the item in the box, closes up the box and applies a label.  The cardboard is supplied in a continuous z-fold. See the video here.

Huge ecommerce vendors like Amazon and Walmart will, I predict, integrate similar box makers at larger scale, and will begin to take ownership of the cardboard waste problem.  Consumers will propel this change, as they increasingly value the ability to choose shipping options that might cost a bit more, and/or take longer, but also minimize the amount of cardboard and packing material they must deal with. Use and disposal of these materials will move from being a residual in algorithms that minimize shipping time and cost, to being a variable to be optimized itself.