Professional Documents
Culture Documents
Demand
How much will customers buy? Try asking them
The News: As reported in the Wall Street Journal, Weyerhauser earlier this year announced a series of
supply chain related strategies and improvements that included plans to present a single face to
customers across formerly different business units, and a leaner, more responsive manufacturing
strategy. (See Weyerhaeuser Restructures to Lower Inventories, Build a More Responsive Supply Chain.)
An outgrowth of these plans, the company is hoping to get a better handle on supply and demand and
smooth out the vicious cycles that have plagued the building products industry. As the Wall Street
Journal reports, Weyerhauser has begun attempting to sell wood products much as retailers such as
Wal-Mart Stores sells consumer goods: Stock inventory based on precise customers forecasts, not
estimates that have often proven wrong.
Now, three supply chain managers roll up data from the companys 43 North American sales regions
based on what customers say they plan to order in the next 6 to 24 months. In the past, Weyerhauser
relied almost exclusively on internal forecasts which surprise! often proved highly inaccurate.
The company then drives the data back up its supply chain. Factory managers meet weekly to compare
inventories and production plans with the forecasts and adjust as needed.
Using this approach, for example, Weyerhauser has been able to keep inventories of soft wood lumber
at about one weeks demand, whereas in the past, with the housing slowdown, they likely would have
grown to 2-3 weeks of inventory.
Like many companies in the wholesale distribution industry, VWR is facing pressure and opportunity on
a number of business fronts. In general, many wholesale distributors are seeing some of their traditional
role and value for the customer being changed by the Internet and other forces. For example, the role of
the wholesale distributor in terms of being the primary source of product information and availability is
very different in a google economy, whether the company is a business-to-consumer (B2C) or
business-to-business (B2B) distributor.
In addition, many wholesalers have been historically successful having a purchasing centric view of the
world, with the company oriented towards buying opportunities and terms, rather than a being more
pull oriented and letting customer demand drive the rest of the companys supply chain and
purchasing patterns.
Pressure from both the buy and sell sides is negatively impacting margins percentages for many
distributors as well.
VWR was strategizing on how to make the transition to being more demand-driven when it was
purchased two years ago by a private equity firm. That added an additional focus on improving cash flow
and working capital both of which are heavily driven by how well a distributor manages its inventory.
Overall, the company has embarked on a broad strategic overhaul, with a significant supply chain and
demand-driven focus. In addition to taking a number of actions to impact cash flow and margins
through better forecasting and inventory management programs, it is also looking to increase its
visibility to its supply chain internally and to use that information to better collaborate on both the buy
and sell sides.
New technology in the area of demand planning, inventory management, and performance
measurement is a key part of that, says Shellenbarger, but must be deployed in the context of significant
process re-engineering.
The company was using a traditional re-order point purchasing and replenishment paradigm that
looked at policies but didnt really consider expected demand. In piloting new technology from supply
chain planning and execution software provider Logility, the company is finding substantial
improvements in inventory levels through more accurately forecasting demand and letting that forecast
combined with trade-off analysis and SKU-based inventory policies determine optimal stocking levels
and purchasing requirements.
A test in one SKU category, for example, demonstrated a 35% reduction in total on-and inventory as a
result of the new processes and software tools. Key to that is using the Logility systems ability to timephase demand forecasts and SKU replenishment requirements, a significant improvement over re-order
point systems.
Like many wholesalers, VWR is also more heavily moving into private labeling and global sourcing, and
these longer supply chains put even more pressure on inventory management. Shorter product
lifecycles and the need to better manage new product introductions and end-of-lifes is also critical to a
distributors. Shellenbarger believes the new planning tools will also help improve results in both those
areas as well.
ADP Brokerage Services Drives Out Cost, Improves Efficiency Though Webbased Portal
ADP runs a 400,000 square foot DC on Long Island that does literature and related fulfillment for many
of the countrys leading brokers and other financial organizations. With rapid growth, the company first
implemented a new WMS (HighJump Software) to help it improve distribution performance, and later
turned to inbound and supplier integration processes to look for additional gains.
Like many companies, ADP was using fairly manual processes for its purchased materials: phone calls,
faxes, email. It generates substantial purchase order and inbound activity (peak volumes of 2500+
pallets received per day) with eight key commodity suppliers.
Beyond overall process inefficiencies stemming from the manual methods, ADP also saw opportunities
to improve logistics processes and costs through improved supplier integration as a result of better
visibility and inbound ASNs.
To achieve these opportunities, the company implemented a web-based supplier integration tool (also
from HighJump). The portal enables the entire purchase order process to be managed over the web,
provides visibility to ADP buyers into vendor inventory positions (ensuring they can buy in truckload
quantities and avoid material stockouts in the DC), and enables suppliers to easily generate inbound
ASNs.
The results: Substantial reduction in receiving times as much as 75%. Receiving processes that used to
take as long as an hour have been reduced to 10-15 minutes through use of the vendor ASNs and
associates receiving processes based on simple scans.
Reduced transportation expense: Visibility into vendor inventories enables ADP buyers to consistently
order in full truckload quantities, without worry that the order will not be full due to supplier out-ofstocks. The system has also reduced the need for expedited freight.
Reduced overhead costs: Significantly reduced time associated with purchase order creation and
tracking automated communication processes. Significant reduction in phone calls/emails, leading to
reduction in total procurement cycle times.
Unified supply chain structure: one supply chain executive with four direct reports one each
for the supply chain processes of plan, source, make and deliver.
Optimization of the physical supply chain: moving some production to lower cost eastern
European countries, use of a new central distribution center that can rapidly ship goods to
smaller, country specific warehouses.
Major infusion of new technology and other IT contributions: normalization of all key supply
chain metadata (customers, items, suppliers, etc.) across regions and development of a
unified data warehouse; new demand planning, production scheduling, and other supply chain
planning tools (Manugistics)
Aggressive use of postponement strategies, making generic bottles of an item in the plant, later
adding customer specific labels and other touches as part of distribution processes.
Much greater collaboration with suppliers, especially around new product design.
Change in some suppliers/sources to increase flexibility and lower inventory, even if per unit
costs were slightly higher.
It wasnt easy. No specific costs were noted, but in addition to the technology and other capital costs,
Avon had a dedicated team of 45 working on the project for 18 months. The result: annual savings
estimated at $50 million, adding two full points of gross margin to the bottom line.
at each step, and required hand-written entries on the traveler that were then key entered into
Seagates MES.
The media is moved throughout this process in a plastic cassette, which holds 25 disks. It is
difficult/impossible to discern visually whether a given cassette has been through a particular process
step or not, problematic because there are many different routings depending on the specific product.
The bar code-based system had many issues, including difficulty quickly quarantining bad parts, missed
or duplicated steps, and a general lack of true process control.
In the late 1990s, Seagate adopted a closed-loop RFID system to manage the WIP track and support
process control. Read-write tags are embedded in the cassettes, and are read before each step in the
manufacturing process. The reader and an associated PLC at each tool are wireless connected to an
RFID controller/server, which in turn is integrated with Seagates MES system and manufacturing
database.
The human element of WIP tracking and process control has therefore been largely eliminated. Tags are
read automatically, and the system ensures that the cassette is at its correct next routing, that a step is
not being duplicated, etc. For some steps, there are also required dwell times, which the system will also
enforce. Bad parts are more easily quarantined, and Seagate managers both in-plant and across the
globe have complete visibility to WIP inventory and the production process in virtual real time.
Seagate is actually writing the production data at each stage to the tag as well. I wondered why this was
necessary, instead of just using an RFID license plate on the cassette and relying on the database to
track the steps, but Mr. Rao said this approach allows Seagate to continue production for up to several
hours in the case the MES system goes down.
Seagate has realized significant benefits from the system: 1-10% yield improvements, reduction in
defects, and complete traceability of product increased from 50% to 0ver 99.5%. Whats next? They are
looking at perhaps using RFID in distribution, but the cost of tags is still too high to justify. They are also
looking at using tags with more memory to capture additional process variables at each step.
forecast accuracy of less than 50%, and we had a perfect storm that was leading to hits on the bottom
line, Dadmun said.
Particularly vexing were the various walls of silence that restricted the flow of information; for
example, little or no information about why the forecast was missed. Wed ask, Why did we miss the
forecast? Dadmun said. Because customers didnt order would be the response. But why didnt they
order? That we didnt know.
There were also walls of silence between the supply side and marketing, and the supply and sales side
and engineering. Only engineering really knew what dates they were going to hit in terms of new
product introductions, Dadmun noted. Often, we were building supply capabilities and revenue plans
that didnt reflect the true schedule.
Faced with this situation, Dadmun helped drive a number of initiatives to improve results. This included
forming a multi-functional team across supply, sales, marketing and engineering to look at the problem,
investing in new supply chain technology, and ultimately starting an effective S&OP process.
There were some key moves. First, to get buy-in, Adtran worked with the vendor (i2) to back-test
forecasting results with its demand planning tool for 20 of the companys top SKUs over the past 2-3
years. The surprising insight: the tools base line forecast simply based on history, etc. was better for 18
of the 20 SKUs than the forecast teams results an observation that quickly generated interest and
support from execs and others.
Adtran also brought in some outside consultants to help them benchmark against best practice in the
area. This also worked to show Adtran execs that they were not operating at close to an industry-leading
level. It delivered a wake-up call, Dadmun noted.
Dadmun also helped convince the organization that while an improved baseline forecast coming out of a
new demand planning tool would help, it wasnt enough. Thats still a rear view mirror view, he
commented. You need additional input from sales, marketing and engineering to see the upcoming
curves and exit ramps.
Phase 1 (complete):
Use of an on-line S&OP dashboard that provides a detailed view of forecasts and actual
results, and allows rapid drill down into supporting information
Focus on identifying when and how the forecast and plan went wrong
Further reduce the walls of silence through both process and technology
Phase 3 (planned):
Facilitate more true root cause analysis, especially in trying to learn at a detailed customer level
why a forecast was inaccurate.
Among the real keys to achieving S&OP success is to really start with the end in mind, Dadmun
noted. You really have to define what you want to come out at the end. This is essential, and
not as easy as you might think.
There is no change without some suffering be prepared to have poor processes and results
exposed to execs and peers
Develop early proof points, perhaps in a pilot mode, to create buy-in
Benchmark to be able to say Ive seen it done.
Use some outside experts/consultants if this is new territory for your company, partner with
others that have been down the path
Work hard to develop other internal champions besides yourself
Celebrate success when one planner hit 72% accuracy, the highest level, they took the whole
team out to dinner
We have to be able to get to the point where we can drill into the demand side of the equation with
the same level of detail and insight that we can on the supply side, Dadmun added.
Just a few years ago, 7-Eleven had virtually no in store systems or strong linkages to corporate systems.
According to Keyes, We didn't know what we sold by store. It was total guesswork. It would take us
almost 20 days to collect sales data from the stores because it was all manual. The chain relied on many
of its vendors to stock stores because they had a much better grasp on what was moving.
Now, 7-Eleven uses bar code scanning, in-store inventory systems, and frequent communication of that
sales data to drive merchandising and replenishment processes. As Keyes notes: We can track every
item literally every hour and use that point-of-sale data to drive revenue. In effect, we have come full
circle in our ability to respond to the needs of the customer. Because now, we can use technology as a
surrogate for being able to talk to every customer who walks in the door. We don't just have to rely on
items we know will sell quickly, like 12-packs of beer; we can actually stock something that we're taking
a chance on. And in short order, we'll understand how the customer responds.
Granted, small C-stores may be easier to manage than large box retailers, but we still believe that
potential improvements using existing technologies such as POS and bar codes to solve many of the
problems being targeted by RFID.
Like many retailers, Penneys had traditionally relied on traditional printed routing guides for its nearly
2000 suppliers. As a result, there were few opportunities for cross-vendor shipment consolidation, and
Penney had to maintain a decent staff just to check whether supplies were actually following routing
guide rules.
Over a couple of years, Penneys has moved to a dynamic inbound process.
When vendors are ready to ship, they send a request for routing (EDI 753 transaction) into Penneys.
Some smaller suppliers are allowed to do this via a web form.
A Penneys preprocessing tool does a variety of validation checks (is it within the shipping window? are
the required data elements present?), after which all the approved requests are sent to i2
transportation planning software. An optimization run is performed at 1:00 a.m., after which an EDI 754
response is sent back to suppliers with routing and pick-up confirmations. Currently, the pick-ups are
scheduled for 48-72 hours from the time of the EDI, to allow Penneys to secure capacity. Shipments are
consolidated among vendors, reducing LTL freight in favor of less expensive truckload shipments, and
often allowing Penneys to bypass its own pool centers entirely and deliver direct to its DCs, which
further reduces transportation costs and takes 1-2 days out of the inbound delivery cycle.
Key to the programs success is that Penney has been able to get support for the program from Penney
execs and merchandisers, especially around the 48-72 hour windows from request to shipment. While
the transportation group accommodates true emergencies when product is absolutely required in less
time, Penneys has been able to maintain this discipline successfully, without which it would be hard to
achieve the results, Persons said.
One challenge was certainly getting suppliers both technology-enabled and to modify their processes to
work with the new model. For example, many lacked cartonization logic, because in the past they could
manually build cartons, then determine routing after the orders had been packed. Under the new
process, most need to use cartonization programs to accurately estimate cartons, weigh and cube prior
to actual order picking processes. Many suppliers also responded by reducing the number of carton sizes
they used, Persons said.
Persons also said that while many suppliers were skeptical of his claims, that the program would actually
enable them to improve processes and get a better understanding of their JC Penney business. Many
have later confirmed these benefits did result from participating in the program, which now controls
about 90% of Penneys inbound freight.
Ingram Books is the worlds largest wholesale distributor of books, managing over 1 million titles and
servicing tens of thousands of customers retail customers.
At the recent CSCMP conference in San Antonio, Don Clayton, VP of Logistics Analysis for Ingram,
described for attendees how an innovative approach to order wave simulation had uncovered key
insights and enabled it to significantly improve fulfillment processes.
Ingrams DC operations have some inherent challenges, given that almost all of the volume is piece
pick based, with same day order-ship policies, and some significant peaks and valleys of order flow. It
also had to manage significantly increased levels of value-added services, and huge increases in
consumer direct shipments. The consumer direct shipments not only drove up the total volume of work
substantially, but significantly drove down the average number of lines per order, which in turn led to
greater use of batch picking and downstream sortation.
As the scope and needs of these customers advanced, however, Ingram found the work in its four U.S.
distribution centers was growing increasingly complex. To manage all the unique fulfillment scenarios,
Ingram had developed a slew of different pick wave types to meet ever more precise customer order
profiles and DC processing needs.
Sound familiar?
According to Clayton, the company decided that exploring simulation of this environment was likely to
result in a number of benefits:
Provide insight into dynamics of a complex operation, and what the real bottlenecks and
constraints were?
Improve staff utilization and scheduling
Ingram hired consulting company TranSystems to develop a simulation tool, which would basically ride
on top of the companys existing, tier 1 WMS.
Get a better handle on what was really happening with order volumes by type, customer,
seasonality, etc.
Understand what the cost and processing times were for different waves types and scenarios
under current practices
Improve those practices based on understanding the data
Simulate what the effect of changes in wave practices and policies would be before actual
implementation
Give wave planners better real-time tools to understand how the wave would flow through the
DC and to better balance the work load for optimal results.
Like many companies, for example, Ingram sometimes found that completing the last few percent of a
wave could take a very long time, leading to operational complexity. The simulation tool could help
them better understand what was causing the end of wave issues, and present new approaches to
mitigate the problem.
The graphic below offers just one example of the kind of insight the tool provided, here showing analysis
of how long a order available for picking during a wave sat between initial release and picker activity,
and how well utilized for the wave pickers would be by pick zone.
Through use of the tool, Ingram was able to make a number of operational improvements to reduce
picking costs and streamline order and material flow.
In some cases, Ingram reduced complexity in the number of wave types it would try manage. The
information highlighted some areas where it DC standards needed improvement, and even identified
some problems in its actual configuration of the WMS (e.g., pick zone definition). The DC had been
primarily optimized around picking efficiency, and according to Clayton the data is showing there may
be benefits in making some trade-offs between picking efficiency and downstream packing and shipping
operations.
Ingram also used the tool to confirm prior to deployment the positive impact of some new picking
strategies in one DC, and significantly improve the actual order wave building and release process by
giving planners better data about how the wave will impact various processing areas.
For the initial implementation of the tool in the companys La Vergne, TN distribution center, data was
collected for about 5 months. For the second DC to go live with the simulator, only 2 months of data was
used before deployment.
Clayton noted that in most cases, wave planners have no real incentive to dig into the issues and try to
drive continuous improvement. While the tool has had a number of benefits, according to Clayton, The
main benefits were to force us to review our DC processes, and to better understand how specific
decisions we made impacted operating results.
Key Insights: It's easy for creeping growth in DC complexity to really start to impact effectiveness
without real insight into how or why. Simply blaming complexity itself doesn't lead to improved results.
However, you do it, get your arms around what the data is telling you.
The as is case was similar to many we hear from companies that have not put a focus on
transportation excellence. As McDermott said, QVCs transportation and inbound processes were
characterized by the following poor practices.
In 2003, we had 12 transportation specialists whose only job was to manually build shipments based on
vendor faxes or emails into our legacy systems, McDermott said.
To be successful, McDermott said, QVC processes have to be built for velocity, built for speed. There
was an obvious disconnect between this principle and its approach to transportation.
As a result, QVC launched an ambitious program that achieved significant results in a relatively short
period of time. The key drivers:
QVC developed a vendor portal that has a robust set of tools for both QVC and its vendors to manage
the PO to delivery process. 100% of our inbound shipments are now built by vendors or forwarders,
McDermott said. There is zero data entry done by QVC.
This includes a series of automated checks to ensure that the shipment meets QVC quality and process
criteria before approval to ship and a routing instruction are delivered electronically to the carrier.
The transformation has had enormous benefits. Transportation headcount has been reduced, mostly
through attrition, and the team now focuses on optimization/planning and working as supply chain
specialists to improve vendor collaboration and integration. Both these tasks add real value to QVC, as
opposed to the low value data entry and administrative work they were doing before.
Work is hugely automated. Transportation specialists get automatic alerts when vendors or carriers
have not completed certain steps when they should have, allowing QVC to move to a more manage by
exception environment.
Visibility has improved dramatically. McDermott said QVCs goal is to get ASNs 96 hours before
shipment. While he acknowledged some vendors believe this is too long a window, but were telling
the vendors we need this, he said. Since QVC plans its television sales programs 5-6 weeks out, it is
critical that supply chain disruptions not force the sales side to have to revamp plans because product
wont be available. We now have high levels of visibility and control, he noted.
The vendors also benefit from the event management technology. For example, a vendor may get an
alert saying it is less than 96 hours until the ship date, and not all of the PO/shipment qualifiers have
been met.
As a result of the transportation and TMS initiative, QVC has also been able to significantly increase the
percent of its inbound shipments that it controls thru collect freight, versus vendor prepaid. It is also
achieving savings on the order of 12% on that collect freight thru optimization and core carrier program
adherence.
There was clearly a lot of change management required both internally and - critically - at the vendors.
QVC also invested in a big educational effort with its vendors to get them up to speed. They relied
heavily on a webinar type format with groups of 20 or so vendors at a time, McDermott said.
Ultimately, McDermott said, QVC is trying to get further inside the vendors door, in terms of the
manufacturing process. We think if we do this right, we can have 3000 supply chain specialists, not just
our internal ones he said, referring to the number of vendors QVC has.
Given the size of its operation, QVCs transportation transformation turnaround over a relatively short
period of time is an impressive story.