Hewlett-Packard is the largest manufacturer of inkjet printers in the world. I was the senior packaging engineer for HP for over 17 years, before leaving and starting my own consulting company in 2005. HP saw great opportunities for selling products into the Indian marketplace and decided in the early 2000s to increase sales by bringing a very low-cost printer to the market, one which had been selling worldwide by the millions since 1999. This particular printer was small, light weight, and had successfully shipped to dozens of countries, along with significant sales through the internet, meaning they had been shipped as single units in very high quantities, and yet suffered an extremely low damage rate everywhere. As a result, it was quite a surprise to find failure rates about 100 times higher in India than the worldwide average. The first thought by many was that material handlers in India must have been dropping this printer from higher drop heights, and be doing it more often, than any other supply chain. Some people suggested that we should change the internal cushioning from moulded EPS (expanded polystyrene, which is the least expensive type of foam), to PE (polyethylene, which costs 2-3 times as much), and perhaps increase the size of the box to accommodate more cushioning. Either increasing the cost of materials or increasing box size would have added up to millions of dollars of extra costs.

Whereas many companies may jump to conclusions and attempt to fix problems without knowing as much as they should, HP took a different approach. Questions that came to mind included:

1. What exactly were the types of damages being experienced?
2. Had we ever seen that type of damage in our laboratory testing?
3. What was the root cause of the damages?
4. If we were to simply believe that more or better cushioning was needed, what laboratory tests should we verify it with?

After all, the current packaging passed our current tests, so it wouldn’t make sense to simply subject new/improved packaging to the same test. Good laboratory testing is able to replicate consistent failures found in the field, but so far, HP hadn’t been provided good data on what the problems actually were.

With the above questions in mind, I set off for India in September 2003. The plan, over a span of two weeks, was to visit a variety of sites in and around Mumbai, Delhi, Lucknow and Bangalore, seeing a total of 22 different kinds of warehouses and distribution centres for HP, distributors, freight forwarders, wholesalers and retailers, along with many stores, and attempt to better understand why some HP products had very little damage while this particular printer had extraordinarily high failure rates. Afterwards, return home, initiate new tests that replicate the consistent failures found, then develop new designs that pass those improved tests, implement the new designs, then return to India in a year to see how the new designs are performing.

What we found
First, the stories were true. This inexpensive inkjet printer had far more apparent damage to the packaging and the product than any other HP product travelling in India and exhibited damages not seen anywhere else in the world. Were our qualification tests lower than other HP product lines? Were our products more fragile than others? Was the packaging really not as robust as that used by other product lines? The answers surprised us…

Anyone who ships products through India should know the following about why it might be one of the harshest supply chains in the world:

Climate: Heat and humidity can be extreme. In June 2003, three months before I arrived, there was a three week span where temperatures hovered between 115 to 125 F in portions of the country. During my stay, it was commonly in the more than 90 F to low triple digits, with 90% humidity. High temperatures and humidity degrade paper-based packaging, such as corrugated boxes, but it also affects the varnish used on the outside of boxes. Many products depend on bicycle rickshaws and open tuk-tuks for the last mile of delivery, even if it’s raining. Are your boxes ready for that?

Dust: We found deposits of dust everywhere throughout distribution, homes and businesses, coating the outside of boxes, but also the inside of products. Dust acts like sandpaper: if boxes are vibrating against each other in the back of a truck, and dust particles are between the boxes, then abrasion occurs. Have your bare products been tested with a coating of dust? I’m a great believer that engineers can design almost anything, but only if they know what they must design for.

Though all HP products arrived into India on pallets, virtually all products were stripped off of pallets once they left the main distribution centre. From then on out, all products were handled one at a time, in every warehouse and every truck.

Without pallets and unitisation, the only way to stack boxes beyond arms’ reach is to step on top of other boxes. Were your boxes ever tested for being stepped on? Our’s weren’t. Also, in depalletising everything, boxes are commonly turned onto their broadest face, even though the corrugated fluting is then in the wrong direction. If the flaps aren’t at the top and bottom of a box that is supporting a compressive load, then the tape holding the flaps shut becomes an important component in surviving compression.

Infrastructure: More developed nations have extensive networks of paved roads in good condition, along with modern trucks. Many countries in Asia, such as China, Malaysia, Thailand, Singapore and others, have improved infrastructures to help commerce be more efficient. China, which built its first 33 miles of freeway in 1987, now has over 70,000 miles of freeway, which is slightly more than the US. In contrast, India has about 800 miles of expressway. Not only are the vast majority of roads in India sub-par, but the trucks lack the softer rides found in places that have better roads. The combination of bad roads, long distances and poor trucks add up to intense levels of vibration not found in most other large economies.

Supply chain taxes: India has a unique system of taxing products that enter large cities, even if those products are ultimately then shifted to some other city to meet customer demand. As a result, the largest distributors and wholesalers have been forced to set up warehouses both at the outer edges of metropolitan areas (to avoid paying a 5-6% tax) and then other warehouses within the city to hold onto fast moving products. Due to the extremely slow roads, let alone the time it takes to pack a truck one box at a time vs a full pallet at a time, they need to have these different piles of products. If the extra tax didn’t exist, there’d be fewer warehouses and fewer handlings. Prior to the first trip to India, the HP supply chain manager provided a supply chain map, which should have been indicative of the number of handlings the products would be exposed to. Only by observing first hand how products moved through the country, did we discover this excise tax and additional movements from one warehouse to another, let alone additional tiers of the supply chain.

Labels: Other than the main HP distribution centres, virtually no one else used electronic scanning to receive products and account for their movements. Instead, people wrote on clip boards what came in, what left, and where it was sent. Every stop along the way would apply a new label on one of the end panels of the box. Is your box ready for lots of labels, thus covering, perhaps, important marketing and product information?

Surprise discoveries
Clearly, the supply chain in India is difficult: extreme levels of temperature, humidity and dust; compression compounded by people commonly stepping on boxes used as step stools; virtually no palletisation/unitisation, thus greatly increasing the number of individual handlings of each product, and having no protection from stretch wrapping that is normally found on unitised loads; road vibration levels that are unparalleled for intensity due to rough roads and poor truck equipment, with boxes bouncing around unfettered, experiencing dramatic impacts from potholes and rough roads. But, even with these difficulties, why did one particular product experience far more damage than others?

By observing first hand how products were being handled, and seeing that some products had virtually no problems while mine had consistent problems, I also noticed one other detail: 100% of the problematic printers came from just one manufacturing site in Thailand. Was that a coincidence? No! It turned out this site was not using materials specified on the engineering drawings. For instance, pallets had fewer deck boards, and used thinner wood, than was specified.

As a result, the corners of boxes fell within the gaps between boards, thus dramatically reducing their compression strength 30-50%. Also, with fewer boards, when a forklift picked up the load, the deck boards bowed upwards, causing boxes to crease mid-panel, from the bottom up, thus further reducing compression strength of the box. We also discovered this factory was using thinner outer liner on boxes than required, along with sub-par varnish. In contrast, HP boxes from other sites used heavier paper weight and more robust, shiny varnish. We conducted testing in our laboratory on Thai-made boxes and successfully replicated the vibration problems seen in India. By establishing the correct vibration input for stacked boxes rubbing on each other, we then made this new test the base line for future products going to India. We also established a rub test for the varnish. We discovered the standard ASTM D-5264 rub test was vastly insufficient to replicate the problems seen on a wide scale basis in India, so we increased the loading and the number of rubs for future varnishes to surpass.

We also discovered that the tape used on these boxes was not what we had specified. Though it looked like tape, the quality was significantly lower than the 3M tape we had expected. The Thai tape broke easily on impact and came loose when the box was turned on its side and compressed, forcing the flaps outward. Once we saw that virtually 100% of boxes are turned onto their broadest face during distribution, we realised we needed to test and design for this condition. Simple improvements to the pallet, tape, varnish and box material completely eliminated all of the problems we had witnessed in India.

Do you remember what was said at the beginning of this article about the initial reaction to damages? Where many thought it must have been because of high drops and more of them? In two weeks of travel in 2003, and then a return trip a year later, I never witnessed even one printer being dropped. I’m sure that drops do occur, but I don’t think it’s a major reason that contributed to the high damage rates seen initially. Ultimately, we did not change the cushioning material or the box size, and yet damage rates dropped to the worldwide average.

Recommendations
Very few companies actually know how their products are handled and what types of damages consistently occur in distribution. Some companies attempt very expensive and time-consuming measurements of the distribution system, believing they can use data acquisition recorders to capture the shock, vibration, temperature and humidity conditions their products are subjected to. These data systems do a very good job of recording events, and especially events that are continuous, but they’re terrible for defining drops, which happen only occasionally, and never record compression, which was a leading cause of failures for the HP products. Even if drop impact was the main problem someone wanted to record, it would take hundreds of product recordings to find the input levels leading to the damages in the field. In contrast, with direct field observation, it becomes rather obvious, in very short order, what the consistent problems are and the root causes of those problems. First-hand observation also allows something very precious: understanding why other products are not failing similar inputs. From my perspective, damage replication in the lab is the fastest and best method for reducing damages and overall costs.

I have travelled all around the world following products through supply chains. No two products are exactly the same, and their supply chains certainly aren’t exactly the same. I had an experience where large screen TVs from Asia were coming to the US. One manufacturer was having far more damages than another, even for the exact same size TVs coming from the same Asian country and into the same distribution centre in the US. How could that happen? It turned out that one arrived on slip-sheets and the other on pallets. My client didn’t have the slip-sheet equipment, so used clamp trucks to pull those large sets off the ocean containers, while the other products on pallets were simply lifted off with forklifts. The clamping onto the screen face caused the damage. My client then ordered push-pull equipment for their distribution centres. Though my client was obviously a part of these warehouses, it was only through my direct observation that the problem was identified. The client had initially assumed the sets must have been lower quality than their competitors’ sets and believed the damages must have happened on the way to the US, but the reality was that they had caused the damages themselves.

Once consistent failures are identified, it’s important to establish laboratory testing that replicates the failures, using the least amount of input possible. For sophisticated electro-mechanical products, I like to establish bare product fragility parameters. Understanding how components fail from shock and vibration can often lead to lower cost solutions than increasing the cost of throwaway protective packaging. Once damages are replicated in the lab, then this new test must become the baseline for future products and packages to survive.

Finally, as seen in the India study, one should verify if all materials are to specification. The poor material used was a total surprise, but in investigating it further, there was an altruistic explanation: people who didn’t know better, were simply attempting to reduce material costs. I love the idea of reducing costs, but doing so without verification testing can lead to extraordinarily high failure costs.
If you distribute products worldwide, and you use only one package design for the entire world, and you have no damage when shipping in India, then this is a clear indication that you are wasting money on excessive packaging for products going to the US, Japan, Europe and perhaps many other places. My printer boxes for India had to increase their compression strength compared to the rest of the world, but sales in India were a small fraction of the overall sales. Increasing costs to survive specific distribution inputs makes sense…it’s cheaper than high damage rates…but it makes no sense to foist those additional costs onto products that don’t need them.