Cost-Cutting Options: Automate or Outsource

By the early 2000s it was clear that American Greetings would have to improve the efficiency of its operations. Steve Willensky, senior vice president and executive sales and marketing officer, recalled, When I began with the company in 2002, the thing that struck me was that our costs were high, we had redundancy in our plants, and the supply chain was a nightmare. American Greetings was considering two approaches to reducing manufacturing costs: streamlining and automating domestic production, and outsourcing to China. It seemed to some in the company that outsourcing most or all production was the obvious solution. But top executives did not want to jump to conclusions without studying the problem. For a number of months the company engaged in an internal debate about the merits of improving its existing systems versus transferring manufacturing overseas.

The Case for Outsourcing

Many senior managers assumed that outsourcing was both desirable and inevitable. As CEO Weiss remembered, Initially there was a bias in favor of sending manufacturing offshore. The reasoning was, Everybody is doing it. We haven't done the math yet, but that's going to be the answer. Indeed, some of the companys competitors had already moved in that direction. For example, in 1998 Gibson Greetings decided to outsource manufacturing of all of its products. Not only was outsourcing a trend; in some ways, it seemed to be a logical extension of what had happened at American Greetings a generation earlier. Originally all manufacturing had been done at company headquarters in Cleveland. But beginning in the 1970s, the company had moved production from the highly unionized industrial Midwest to Kentucky and Tennessee. This brought the advantages not only of cheaper labor and higher productivity but also lower transportation costs, because manufacturing was now done in the center of the country. At first glance, outsourcing seemed to be a similar decision. As Goulder recalled, Many long-time company executives drew an analogy and suggested that the outsourcing decision was the same as when we went to the mid-south. American Greetings was already sourcing gift bags from Asia, so why not greeting cards? The greatest advantage to outsourcing seemed to be the expected reduction in labor costs. Gary Kopechek, vice president of global sourcing and services, pointed out that because every card was unique, the business involved not only variable labor but a high amount of overhead, as well. We get huge seasonal spikesChristmas, Valentine's Day, Mother's Dayon top of our everyday business, which is relatively flat, said Kopechek. That's problematic in keeping our work force fully utilized. Outsourcing to vendors allows them not only to produce our high labor content cards but also to schedule our seasonal spikes into their overall production mix. Outsourcing manufacturing to China or Mexico seemed to offer a significant labor advantage. The initial design and planning phases needed to remain local, but manufacturing could be done anywhere.

The Case for Automation

Other executives, especially those on the manufacturing side of the company, believed that they could best maintain the quality and efficiency of operations by streamlining manufacturing in the United States and investing in upgraded equipment. These executives pointed to problems with transportation, lead times, and training that would occur if production were sent to a distant location. In particular, the logistical costs of outsourcing were significant. To ship product halfway around the world was more expensive than to truck it from the mid-south to the companys distribution center. The problems caused by the longer distances were compounded by the lead times required for some cards. With manufacturing located in the mid-south, maximum flexibility could be given to the companys planners and designers. It was relatively easy to send last-minute changes to the plant or to add rush jobs for important customers. But with the increased transportation times that would come with outsourcing, this responsiveness might be degraded. Chris Furlong, vice president of supply chain planning, noted that seasonal cards were particularly hard to send offshore because of the tight schedule required at the specification and design phases of production. Let's say creative has to output 1,000 new designs for Christmas, explained Furlong. They cannot process enough of them early enough to get them to China and back. In considering a new supply-chain strategy, it would be essential either to do some short-term production close to home or to pay high costs for air-freight for the inevitable rush jobs. As sales and marketing head Willensky put it, Working with a foreign vendor is an unforgiving system. You cant call your local plant down the street and say, Get that job done tomorrow! You are dependent on people far away, and more can go wrong. Another challenge that had to be considered was the difficulty of training a foreign manufacturer. Because American Greetings was used to working with domestic plants that had been in the business a long time and knew card manufacturing, it had relatively rough and informal specification documents. But to potential vendors in China, mass-production of greeting cards was new. Chinese manufacturers had the capability for high-speed printing of catalogs and similar products, but they had little experience with greeting cards. This meant that it would be crucial for American Greetings to specify its processes precisely, explaining how to lay out, print, and finish all of the variety of cards it created. It would take time for Chinese manufacturers to update their equipment and processes to be able to handle high production volumes efficiently. The sheet arrangement was the most intricate problem for a new manufacturer to solve but not the only one. It was also a delicate matter for a card manufacturer to match the colors and designs of the card template. Finlay said, If we were going to outsource, we wanted the vendor to develop the capability to print to our standards. We are accustomed to sending the manufacturer a file that contains the information the way we want it to appear, and its up to them to get it on the paper that way. A parallel problem was that a foreign vendor who did not understand card production would also not know how to estimate costs. Greeting cards are complex in terms of pricing, said Robert Dedinsky, director of global sourcing and services. Because other products are one-up operations, it's easier to get a quote for a manufacturing job. But cards are based on sheeting efficiencies and on variations in the printing and finishing operations required, so the price will differ depending on the pool of cards the vendors get. Manufacturing executives and others at the company believed that it was important not to underestimate these communication challenges. As global sourcing head Kopechek put it, Its easy to sit here in Cleveland and say, Buy me a party favor in China. But there is a lot of instruction that we have to give the suppliers so they know how to make the cards, what testing they have to perform, and what tolerances we'll accept or not accept. American Greetings was especially cautious because of the struggles that its former rival Gibson had encountered in outsourcing. CEO Weiss said, When Gibson Greetings decided to outsource their whole greeting card business, they had a hard time shipping their seasonal cards on time. Because the first Christmas they went through was difficult, the company lost confidence, and the retailers lost confidence. This was a good company with good people. So we were very sober about the idea of outsourcing. We didn't want to fall into the same problems.

Outsourcing by Card Category

Goulder then requested that detailed cost comparisons of outsourcing versus domestic automation be performed for all cards, both everyday and seasonal. These analyses were to be cut for each of the four main product categories: 1) simple ink-on-paper cards; 2) moderately complex cards; 3) very complex cards without hand finishing; 4) very complex cards with hand finishing. These four types of cards required progressively greater labor to produce, the last having the highest labor content. In performing the cost-benefit analysis on the four categories, the base case was the current production in the U.S. before any improvements in process or technology. (See Exhibit 5 for annual card volumes as well as labor, material, and allocated overhead costs for each of the four categories in the base case.)

Outsourcing for Category 4 Cards

The Card Supply Strategy Team first evaluated the economics of outsourcing cards in category 4. These were the most complex items that involved finishing operations that could best be done by hand, such as adding ribbons and other decorative attachments. Because of the immense variety of finishings required, investing in more automated technology was not a practical option for category 4 cards. After discussions with companies in China, the team estimated that the Chinese bid price for a category 4 card would be 15 cents per card. The Chinese bid price reflected the fact that in the near term at least, productivity was expected to be much lower in Chinese plants compared to production in the U.S. This was true for category 4 production as well as for production of cards in categories 1 3. (See Exhibit 6 for wage rates in the United States and China. See Exhibit 7 for a breakdown of labor costs for printing and hand finishing for U.S. manufacturing of category 4 cards.)

Transportation, Lead Time, and Inventory Considerations

After coming up with a basic per-card cost for outsourcing category 4 cards, the team wanted to factor in the impact of transportation costs, lead time, and inventory costs for outsourcing any of the categories of cards. The costs of shipping from China were straightforward. (See Exhibit 8 for transportation costs from China to the U.S.) The impact of outsourcing on lead time and inventory was more difficult to calculate. Total card production consisted of seasonal cards and everyday cards, with total volume across the four categories divided evenly between the two types. Outsourcing to China would only affect the inventory levels for everyday cards. Order quantities for seasonal cards were set well before the start of each seasonal campaign, and Shefrin believed the quantities would remain the same whether the cards were produced in China or in the U.S. Printing and hand finishing category 4 cards in China would add eight weeks to the average lead time. There was concern throughout the organization, particularly among manufacturing managers, about the impact on safety stock resulting from the much longer lead time required for product outsourced to China. For everyday cards American Greetings used an economic order quantity with uncertainty (Q,r) model to determine its order quantities and reorder points. For each SKU, it set the reorder point and therefore the safety stock by specifying a fill rate of 98 percent. In assessing the impact on safety stock, Shefrin treated the six-week domestic production lead time as constant, but he recognized that the estimated 14-week lead time from China was an average and that actual lead time might be longer because of bad weather, port problems, or other delays. He had no data on actual lead times from China and therefore had to make an estimate of variability based on judgment alone. He assumed that the standard deviation of the lead time from China would be about four weeks. In order to compensate for the variable lead time, Shefrin decided to add one standard deviation to the mean lead time. In other words, the lead time could be treated as a constant 18 weeks. Shefrin wanted to compare the cost of carrying safety stock for everyday items under the China option to the cost of carrying safety stock for the current system of producing category 4 cards in the U.S. He found that there were 1,571 everyday SKUs in category 4. With 50 percent of total volume comprising everyday cards, he determined that the annual average demand per card was about was 60,000, which was 5,000 per month or 5,000/4 = 1,250 per week. The company used an order quantity Q equivalent to six weeks of demand or 7,500 cards, and the firms inventory carrying percentage was 10 percent. From historical data Shefrin estimated that the standard deviation of weekly demand for this average SKU was 600. (See Exhibit 9 for a summary of inventory-related data). Assuming demands from week to week were independent, he had enough information to calculate the safety stock for the average SKU under the China option and the safety stock under the U.S. option.

Automation versus Outsourcing for the Other Three Categories

Finally, for categories 1 to 3, Shefrin and the team examined the costs and benefits of investing in new technology compared to outsourcing to China. First the team developed a factory-of-the-future plan that called for an investment of about $30 million across all card categories. Compared to current manufacturing in the U.S., the analysis showed an internal rate of return of about 40 percent and a payback period of about two years. The main effect of these investments would be to reduce the number of labor hours required for domestic production. See Exhibit 10 for the manufacturing costs for categories 1-3 if the factory-of-the-future technology plan were implemented. Second, after discussions with Chinese factory owners, Shefrin estimated bid prices for cards in categories 1-3 at the factory in China. For categories 1, 2, and 3 respectively, these were 3.6 cents, 5.9 cents, and 9.8 cents per card.

1. How do the "net" costs per card compare for each card category:
   • Current structure vs.
   • Outsourced to China vs.
   • Factory of the future
2. Whats the lowest cost alternative for each card category?
3. What combination of alternatives provides the total lowest cost for AG?
4. How much can they save in total (annually in millions of dollars)?
5. What are the strategic pros and cons to be considered before outsourcing?

Notes/assumptions:

• • Shipping costs are cents per card (for all categories)
  • Ignore inventory holding costs
  • Assume the Factory of the Future investment is evenly amortized on all applicable units over 15 years

• Using the information provided in the case, calculate the actual shipping costs (in fractions of a cent on a per-card basis)
• How does this impact your answer to #4 above (annual savings)?

\begin{tabular}{|c|c|c|c|c|} \cline { 2 - 5 } & 1 - Simple & 2-ModeratelyComplex & 3-VeryComplexwithoutHandFinishing & 4-VeryComplexwithHandFinish-ing \\ \hline Category & Cost per Card & Cost per Card & Cost per Card & Cost per Card \\ \hline Indirect Labor & $0.002 & $0.004 & $0.050 & $0.077 \\ \hline Materials & $0.011 & $0.022 & $0.006 & $0.006 \\ \hline Allocated Overhead & $0.008 & $0.015 & $0.046 & $0.046 \\ \hline Total & $0.033 & $0.066 & $0.004 & $0.059 \\ \hline \end{tabular} Exhibit 6: Wage Rates in U.S. Plants and China Exhibit 7: Direct Labor Costs for U.S. Production of Category 4 Cards (Very Complex with Hand Finishing) Exhibit 8: Transportation Costs from China Transportation Conversions: 4 cards =1 unit 72 units =1 "C" Box 1"C Box =23 x 8.5 x 7.5"=0.85 cubic feet Exhibit 10: Labor Costs by Category for Factory of the Future