The Operating Chief #3: Every Box You Ship Is Half Empty 📦

Welcome to The Operating Chief, where we go behind the scenes with the operators, logistics leads, and fulfillment architects who keep supply chains moving every single day.

In this edition, we have James Malley, CEO and Co-Founder of Paccurate. James grew up watching his father build the first multi-carrier shipping system, detoured into theater school at NYU, and spent the last decade processing over a billion packing decisions for some of the most complex fulfillment operations in the country.

In this conversation, James talks about why bad box choices are quietly draining margins across the industry, how new carrier surcharge rules from FedEx and UPS have made packing a CFO-level conversation, and why the smartest thing a supply chain operator can do right now is take the air out of the box.

Your father, Bob Malley, built the first multi-carrier shipping system. You grew up in a household where parcel software was the family business, then went to theater school at NYU and studied at RISD. Walk me through that arc. How does someone go from stage direction to corrugated cardboard, and what did you learn about parcel shipping that others often miss?

I thought shipping technology was mildly interesting as a kid, but my passion was theater, and I was lucky to have extremely supportive parents who encouraged me to try it. In college, I ended up spending a lot of time making websites and applications, and by the time I graduated, I knew that’s where I wanted to spend my time professionally.

After graduation, I reconnected with an old high school buddy who had become a software engineer, and he’s been my business partner and co-founder across multiple companies ever since.

Once we started working on supply chain tech, I diverged from the “family business” that focused on optimizing transportation and quickly found operations and automation inside the four walls fascinating. It’s been over a decade now that our sole focus has been leveraging packing logic to influence robotics, labor, and waste reduction. 

You have described your fascination with what you call the 'butterfly effect' of packing, that a 10% smaller box cascades into 10% fewer pallets, 10% fewer truck trips, and measurably lower emissions. After processing over a billion packing decisions, what is the most surprising downstream consequence of a bad box choice that you have observed in the data?

I think this is why a lot of logtech people never leave the industry after getting into it; the economies of scale we’re dealing with mean tiny tweaks have big network-wide consequences. And it’s quantifiable. There’s something very satisfying and addictive about that.

You have talked about visiting fulfillment centers where the packing logic is governed by post-it notes on the wall and what you call 'tribal knowledge,' the experienced packer who just knows which box to use. Describe the worst case of this you have encountered. What did the operation look like, and what was it costing them that they could not see?

Tribal knowledge isn’t necessarily a bad thing unless you have a high turnover rate. Where we see shippers getting into trouble most often is by ignoring packing SOPs. Over time, packers trying to be helpful will solve problems with “process” rather than bothering operations leaders about it. As a result, it’s extremely rare for a VP of Operations to know what’s actually happening at their pack stations. 

A recent example: an ops leader was trying to figure out why they were going through so much fill material compared to prior years. When she actually went to watch the pack stations, she found that packers were putting an extra layer of bubble wrap on the bottom of every box just because a year prior, a single SKU was getting consistently damaged. Their process fix was well-intentioned brute force, and it resulted in a tremendous amount of waste.

The average e-commerce package reportedly contains about 40% void space. But that is an average. In your data, what does the distribution actually look like? Are there categories of products or types of operations where the void problem is dramatically worse than others? And conversely, has anyone actually solved this to your satisfaction?

We track these numbers by industry vertical, and there’s some variance between auto parts, apparel, electronics, etc., but the most dramatic gaps are between big and small operations. Lower-volume facilities tend to have emptier boxes than high-speed facilities, which is ironic because typically smaller shippers are hurt by shipping air more than large shippers that have more negotiating leverage with their carriers. Big shippers that have Pack-size machines tend to have the most efficient packaging.

Among SMBs, the exception is savvy D2C brands like Our Place that care a lot about “the unboxing experience.” They often have specialized shipper boxes custom-cut for their flagship SKUs and a small collection of boxes/mailers for multi-line shipments. They’ve put a lot of thought and care into it because the customer experience is central to their strategy.

Both FedEx and UPS started rounding up every fractional inch on August 18, 2025, and both introduced cubic volume surcharge thresholds in early 2026 (10,368 cubic inches for Additional Handling, 17,280 for Large Package/Oversize).

FedEx now enforces a 40-pound minimum billable weight when dimension-based surcharges are triggered. For a supply chain operator reading this who manages carrier contracts, what is the single most expensive mistake they are probably making right now in how they respond to these changes, and what should they do instead?

For transportation managers, the biggest mistake is thinking they can fix the issue with sheer negotiating strategy. These are not rate hikes; they’re rule changes. The time of shrewd negotiation being enough is long past.

That means you need to engage the ops and packaging teams inside your organization to solve the problem with you. Ops folks have traditionally been laser-focused on labor savings as a north star, but with these rule changes, transportation costs suddenly become the priority a lot of the time. You might need to recruit your CFO to drive it cross-functionally.

You have a patent on cost-aware cartonization, and you have built what you are calling a Packing Control System. But I want to push the limits. What does cartonization software not solve? If I implement perfect software-driven box selection tomorrow, what percentage of my total packing waste and cost problem have I actually addressed, and what is left on the table?

In pursuit of “perfect packing,” we just keep building the tools we think are necessary to address the problem holistically, mostly based on customer feedback. So the Packing Control System is a collection of tools that interact with each other in a loop:

1. PacAPI-  The Cost-Aware Cartonization Engine
2. PacHealth- Monitor performance and get insights from your data
3. PacSimulate- Run “what if” scenarios and determine what boxes, mailers, bags, or automation you need
4. PacManage- No-code interface for managing packing rules/SOPs/costs/etc.

PacManage has gotten the most fanfare this year; so much of the packing/packaging logic is buried in hard-coded integrations done by people no longer working at these companies. Being able to manage all of it in a single, easy-to-use UI has been really popular, especially among 3PLs who are trying to follow the packing directions and handle the edge cases for hundreds of brands they fulfill for.

Your software generates 3D packing instructions and visual diagrams for packers. But warehouse operators tell us that packers often ignore system recommendations when they are under time pressure; they grab the nearest box and move on. How real is the compliance problem, and what have you learned about changing human behavior on the pack line?

Every company is different, which is what keeps it interesting. Probably half our customers use the 3D visual: it can be especially helpful if their costs incentivize very dense packing or they’ve got fragility or compliance constraints. For the other half, they may just show the recommended box/mailer name on a screen or on a print-out, or have a light pop-up under the correct box size in the rack.

Where it can be really helpful is in cobot pick-to-box scenarios- showing the picker where things go in the box (on a screen on the robot) can ensure the pick path doesn’t get slowed down.

Typical WMS cartonization is often ignored because the logic goes stale and starts making insane recommendations.

Amazon built something called the Package Decision Engine, an AI model using deep learning, NLP, and computer vision that has reportedly helped avoid over two million tons of packaging since 2015. They train it on millions of successful and damaged deliveries. Your approach is algorithmic and cost-aware rather. How do these two approaches differ? How can a small and medium-sized shipper get access to the same packing intelligence with Paccurate?

Amazon is special. They own their own distribution and now offer it as a service, so they’re optimizing for different things than we are for even our largest Fortune 500 customers. And the sheer scale they’re dealing with means they can make dramatic Scope 3 emissions reductions with policy changes like, “Ok, now we’re not overboxing boxes of diapers anymore. Just slap a label on it.”

AI is good for uncovering highly effective policy changes like that, or for identifying item characteristics. AI is not good for actual cartonization- its accuracy isn’t too much worse than deterministic algorithms, but in our tests it’s about 1,800X slower and many orders of magnitude more expensive. It doesn’t really make sense to burn a swimming pool full of water every day to get similar results much more slowly.

Interestingly, Amazon hired a bunch of PhDs to build a toolset at least thematically similar to the Paccurate PCS (Execute, Monitor, Simulate, Manage), but from my understanding, those tools didn’t enjoy wide adoption globally. 

Carrier rate cards are structured to incentivize certain packing behaviors. You have said that sometimes it is actually cheaper to ship two boxes instead of one, or to use a slightly larger box, because of how rate breakpoints work. For our readers who negotiate carrier contracts but have never thought about packing as a rate optimization lever, explain how that counterintuitive math works. A concrete example with real numbers would go a long way.

Sure. In a rate table, each weight has a base price for each zone and a per-pound $ increase. Those costs don’t scale linearly across the rate table. As a rule of thumb, the further you ship something, the more the extra pounds (or air in the box, because to a carrier, dimensional weight is still weight in pounds) will be penalized. As a result, you could have a 3-item order that makes sense to ship in one box if it’s going to the next town, but packed much tighter in more boxes if it’s going across the country where capacity is constrained. If it’s going on an airplane, they penalize air most severely.

If you look at it from the opposite side, the carrier may look at a ton of perfectly packed boxes and say, “This is great for saving total space, but this is harder for a delivery driver to handle, so we’ll incentivize fewer boxes in this part of your rate table.”

Here’s an example we looked at for one of our white papers: two identical shipments, one going to Zone 2 and one to Zone 8. For Zone 2, the price increases per pound were relatively small. In Zone 8, the per-pound price increases at 2.8 times the rate as in Zone 2; hence, the different optimal packing solution.

We can prove it by swapping these pack solutions: If you take the packing solution for zone 2 above and rate it for zone 8, you would lose about $1.10. And if you took the packing solution for zone 8 and rated it for zone 2, you would lose a much more substantial $8.07.

Things get really spicy when you add in materials costs, picking/packing labor costs, and fuel surcharges. A newer customer of ours found 30% labor savings from injecting our engine into their G2P system wave plan, and was willing to accept an average $3 transportation cost increase per box to achieve it. Every shipper is different, and their costs are constantly changing.

But for someone reading this, the simplest way to get started is to ask yourself how you would negotiate differently if you didn’t care what your dim factor was. Taking the air out of the box gives you leverage.

You recently announced the Packing Control System as a new category of fulfillment software, alongside WMS, WES, and WCS. Category creation is an ambitious bet. Make the case: why does packing need its own dedicated system? And then the harder question: what has to be true about a company's operation before a dedicated packing system delivers more value than just upgrading the cartonization module in their existing WMS?

The difference between a cool niche tool and a whole new category is actually more about what’s going on in the world than the solution itself. Right now there’s a confluence of intense pressures on packing/packaging:

Gen-z is far less accepting of waste, and will churn if you’re shipping them air or too much fill material

The companies that see these things and decide they need a real packaging strategy often realize they need a system that’s laser-focused on the problem. It’s too deep and complicated a problem for WMS/WES/TMS systems to solve on their own.

The best validation we’ve seen is other companies popping up copying the language that we had to invent for this category, like “cost-awareness.” Often they will copy our brand colors and other things that we don’t love, but ultimately it’s a good thing. You can’t be a category leader if you’re the only one in the category.

Your exclusive partnership with Packsize puts your software inside their box-making machines. But you have also said that on-demand box-making machines are not smart enough on their own to make the right cost decisions. For a company evaluating whether to invest in box-on-demand hardware, what should they know about the gap between what the machine can do and what the software layer adds, and at what daily volume does the combined investment start to justify itself?

Packsize is a great partner, and I wouldn’t say on-demand box machines aren’t smart without Paccurate. I’d say we help them achieve greater potential. Most box-first machines (ones that make boxes before packing rather than wrap corrugated around items) are handicapped by WMS cartonization. WMS’s can only handle an average of 500 potential box sizes. Maybe 900 if you’re lucky. We dynamically create box sizes of infinite variation, within the tolerances of the specific machine. So, on that alone, we can help pump up the ROI on the machine investment.

For box-last machines, like Sparck, which Packsize recently acquired, they’re usually not handling all the volume in the building. Where Paccurate comes in is at the decision point upstream. For every order, the PCS will decide if an order should go to a traditional pack station, an autobagger, a box-last machine, or somewhere else.

Where we really shine is if a shipper is trying to hook on-demand packaging up to other automation systems. ASRS needs to know what sequence to bring the items out in to make the perfect box. AMRs need boxes that are optimized for their platform/shelf space, on top of the other costs the shipper may have. Heavily automated environments desperately need a Packing Control System to stitch everything together and manage it.

Extended Producer Responsibility laws are now live or approaching enforcement in seven U.S. states. Oregon collected its first packaging EPR fees in July 2025. California's fees are expected in late 2026. For a supply chain leader who has heard about EPR but has not yet had to deal with it operationally, what specifically should they be doing now to prepare, and how does packing efficiency connect to their future EPR compliance costs?

It’s time to have a cross-functional pow-wow if you haven’t already. Bring the packaging engineers and the transportation managers together to really understand how these laws will affect your business. These laws vary greatly from state to state, so you may have to make some decisions that will affect fulfillment processes. “Here’s my pack station just for California, which requires this expensive tape,” is a sentence someone will say out loud pretty soon.

You grew up watching your father build the first multi-carrier shipping system. You are now building the packing intelligence layer on top of the shipping infrastructure his generation helped create. If you could fast-forward ten years, what would packing look like in 2036, and what would it mean for the 50,000-plus warehouses operating in the U.S. today?

I’ve heard the concept of a “dark warehouse” ever since I entered the industry 15 or so years ago and used to chuckle at the idea. It finally feels like it’s just around the corner as a mainstream possibility. The leaps in AI and robotics over the last 6 months are astounding. Paccurate already drives robots today, but now we’re working on making it so AI agents can sign up for an account without human intervention.

I can imagine in 10 years that you drop a robot in a warehouse and it thinks, “Ok, I know what I have to fulfill today, let me grab the right software helpers wirelessly, and get to work. No need to tell the pesky humans how I’m doing it.”

Final question. You have processed a billion packing decisions and seen inside the operations of companies ranging from home goods retailers to cold chain meal kits. What is the single most counterintuitive thing you now know about how things get shipped in this country that the average supply chain professional would find genuinely surprising?

I’m not sure how counterintuitive this is, but I don’t think most supply chain folks intuitively understand the role they personally play in keeping humanity going.

Supply chain is the greatest invention humanity has made. It’s the 8th wonder of the world. It underpins all of our lives from food to medicine to… well, literally everything. The global logistics market is worth $10 trillion. The value of all goods moved through the supply chain is worth $33 trillion, give or take. So much treasure and sweat are expended just to keep it moving.

Hold that idea in your head, and now also accept that every box, truck, and ocean container is on average… half empty. Half!

Imagine the impact on cost and the environment if we doubled spatial efficiency. It would take dozens of vendors like Paccurate and a level of routing logic and international (public and private) cooperation that’s incomprehensible today. But it is a beautiful goal to aspire to, and every supply chain professional will make multiple decisions in their career that will affect how quickly we move toward it collectively.

If you enjoyed this conversation, please recommend us to your colleagues and friends. More exclusive content and interviews from across the supply chain world are on the way.