London Drugs is the fourth-largest pharmacy chain in Canada, but it is hardly a traditional drugstore. In additional to healthcare and personal care products, this privately held retailer also sells general merchandise, computers and home entertainment equipment. Founded and based in British Columbia, most of London Drug's 70 stores are in the western provinces, but the company is expanding eastward as it grows. To evaluate the impact of this shift on distribution, management undertook a network optimization project, which Mahlman headed.

Q: How did you become involved in supply chain activities at London Drugs?
Mahlman: I am a pretty typical retailer. I started as a part-time stock guy with the company 21 years ago, while I was going to post-secondary school in Vancouver. I continued my career with London Drugs in a number of capacities, including store management and district management, basically working my way up. I came to the head office to coordinate some business process improvement and re-engineering work and later moved to Human Resources. While I was vice president of Human Resources, I took on the additional responsibility for distribution and warehousing. My current title is vice president of retail operations and distribution. I also am vice president of TLD Computers Inc. and of Sonora Resort, which are sister companies within the H.Y. Louie group of companies. London Drugs is 60 years old and our parent company is 102 years old. We are one of Canada's largest privately held retailers.

Q: Can you give us a profile of London Drug's supply chain?
Mahlman: We have approximately 70 stores in four provinces in western Canada. We have one main distribution center in Richmond, British Colombia, just outside Vancouver, which is about 450,000 square feet. Another smaller, third-party facility of about 70,000 square feet is in Edmonton, Alberta. And we just brought on line about 250,000 square feet of leased space.

One of the things that makes London Drugs unique is that, in addition to typical drugstore products, we also sell pretty expensive home theater systems, complete with the ability to architecturally install them in homes. We also are one of Canada's largest retailers of computers, one of the country's largest retailers of big-screen TVs. We also sell housewares, hardware and automotive products. And we are one of Canada's largest photo finishers.

Q: How did you develop such a varied product base?
Mahlman: Our heritage actually comes from the photo side of the business. When we started 60 years ago, one of the things that made us unique was that we sold very high-end cameras. Today we continue to be one of the premier photo and digital imaging dealers in Canada. We have fewer stores than some of the other chains, but our typical store footprint is around 30,000 square feet, which is significantly larger than a typical drugstore. Before Wal-Mart came to Canada, we were regarded as the Wal-Mart of the North. The biggest difference between what you see in a Wal-Mart versus what you see in London Drugs is that the quality of our merchandise typically is a little higher. For example, we sell $2500 printers in our computer department-things you just don't see in general retail environments. So we are very unique.

Q: How many SKUs does a store typically carry?
Mahlman: Anywhere from 45,000 to 50,000 SKUs are active in the store at any one time. The size and variety of our inventory certainly presents some interesting logistics and distribution challenges. And we import quite a bit of merchandise. Having our head office just outside of Vancouver gives us good access to Asia.

Q: What are operations like at your main distribution center in Richmond, BC?
Mahlman: What makes Richmond unusual is the wide variety of merchandise that we handle there, everything from classified pharmaceuticals to big screen TVs to extremely high-end and high-value electronic items.

Within the walls, we have regular pallets, we have split-case product and we have hand picked items. Some products are picked to conveyer. Where that is not appropriate, we use pick-to-light technology and we also use carousels. And we have a wide variety of material handling equipment.

About 80 percent of our stores are replenished from our distribution network, with the balance coming direct to stores from vendors. The frequency of deliveries varies, depending on the velocity that the store requires and the logistics of getting shipments there. For example, we have stores that are within a kilometer of our DC, while others are 3,000 kilometers away. And their demand profiles are completely different.

We also trade in cities where there are noise laws, so the size of the truck and when it can deliver all have to be matched with the demand profile. The lowest number of orders a store would get is two a week and we have some stores that get six orders a week.

Q: Why did you decide to do a network optimization study?
Mahlman: We are based on the West Coast and as we grow and expand eastward, the length of the supply chain to our stores is becoming longer and longer. Canada has huge geographical challenges-great distances with low population density between cities, which is very different from the U.S. model.

We already had inventory in six different outside storage locations at our peak season, so managing the refill to the DC for pick was and is a monumental cost and challenge that taxes our systems.

Also, we are looking to replace our warehouse management system and, before doing that, we wanted to be very clear as to what our physical requirements were going to be in terms of number of buildings and the type of material handling equipment and automation equipment that would have to be driven from the system. Those are the factors that drove us.

Q: Were you thinking of adding another DC?
Mahlman: Yes, we had clearly run out of room in our primary DC in Richmond (near Vancouver) and we needed to go through that decision tree of whether it would be better to increase the size of that facility or to have multiple DCs. One of our challenges was that Vancouver is one of the most expensive real-estate cities in North America, so real estate is at a premium. Yet we knew that in other cities in Canada, real estate is relatively inexpensive. And while our concentration of stores is higher around Vancouver than anywhere else, we wanted to look at the cost of real estate, the cost of constructing a distribution center and the cost of maintenance. Insurance was also a factor, because insurance costs continue to rise, as well as delivery time to the stores.

From an executive perspective, our biggest concern is that the more DCs you have, in theory, the more safety stock will creep into the system and you'll end up carrying unnecessary inventory. Like any retailer, we operate on extremely thin margins, so inventory carrying costs and inventory turnover are key metrics in terms of profitable operations.

In the process of doing some earlier "blue sky" work around our WMS requirements, we had talked to a number of vendors and had initiated a partnership with ESYNC. In getting to know ESYNC and what they could and couldn't do, it was very clear to us that they had a lot of competency in modeling and helping with the types of pain points that I just described.

Being a private company, we don't trust easily or share information easily, so it was important that we find a vendor that we were comfortable with, and we felt that ESYNC understood these issues.

So we went through a voyage of discovery where we established the parameters of the study and gathered information on our demand and order profile over the past three years. We also looked at transportation costs and other costs and all that was loaded into the Insight network optimization software. It produced various solutions that we then discussed as an executive committee. We tweaked and massaged these to where we now have a high-level plan of what we expect to do in the next five to six years.

Q: Was it difficult to get together the information required?
Mahlman: It was, but that was mainly a London Drugs issue. We had never taken on a task with this level of minutia before. We certainly looked at these things from a general planning perspective, but not having to load in actual transaction data over a long period of time. Because of the velocity of our stores and the velocity through the DC, the volume of data was staggering. We had to go through a process of figuring out exactly what data elements the software needed to work best and match that to our historical data and then reformat it in such a way that the software could utilize it.

It probably took us a couple of months to actually get through this iterative process and to cleanse the data to make sure it was error free and that it actually presented the information that we thought it did.
To ensure that we would be confident in the predictive numbers that the model came up with, we took actual numbers from a couple of years ago and ran them through the software, then compared the results to what actually happened in the following year to see the level of accuracy. That also allowed the software to be adjusted to make sure that it was capturing any uniqueness to our demand profiles and to generally make sure that it was accurate.

One of the biggest challenges that I always worry about with modeling software is that unless you have an accurate baseline measurement to indicate how the algorithms are working, you often are never certain of how accurate it is until after you have made your decision. One thing that ESYNC was extremely good at was making sure that we were all very comfortable with the modeling software to begin with and with what it could do, as opposed to jumping right to what the software was going to tell us.

Q: What did the results tell you?
Mahlman: Generally speaking, it verified for us that a second DC was required now and certainly within the next few years to meet our demand profile. And that it made sense to locate that DC in the province of Alberta, where we have our second-largest concentration of stores. We had modeled locating the DC in Ontario, thinking that we could intercept product being shipped out of the east and do a cross-dock there. We looked at multiple, multiple scenarios-putting it at Winnipeg, which is sort of in the middle of Canada, all sorts of things. But at the end of the day, it made sense for us to build in Alberta and to look at expanding our current facility.

Also, it indicated that we should continue to use our current facility for all less-than-full-case picking, because of the absolute dollar investment required for that specialty handling equipment. And it showed that we should do full case picking out of Alberta, so that is something that we are planning to do.

We also looked at different transportation options and at different growth scenarios. If we were to open 10 stores in a certain period of time in this area, what would happen? If we were to expand at a different rate, what would happen? If we changed out full case vs. split case profile, or if we continued to go down certain paths, what that would do to building design and demand-we looked at all sorts of things.

So through the modeling it proved and disproved a number of our cost assumptions and a number of our assumptions about where would be the most logical place to build a distribution center.

Q: Where do you go from here with the plan?
Mahlman: We are starting to do more homework on the design of our facilities. We are taking it to the next level and doing some profiling of our current SKU base to tell us exactly which SKUs we should house in which facility and at a more precise level. We also have looked at some of the upstream or downstream issues, depending on your perspective, such as reconfiguring our purchasing systems.

As part of our ongoing work with ESYNC, we have made a number of process changes-things that we identified as areas where we could be doing a better job within our building. Our goal, before we picked a warehouse management system, was to make sure that we knew where we were going and that we were as efficient as we could be, before we tried to match a software to our needs. Then and only then, when we had all those pieces in place, would we start to look at software options. That way we could match where we are going as opposed to where we have been. So I've been busy with that and we are just about ready to begin the evaluation process.

Q: Were there any results that surprised you?
Mahlman: We thought that the percentage of goods coming direct from vendors to our stores was higher than actually was the case. Having fresh data in front of us to that effect was a surprise. Overall, the process assured us that the vast majority of our operations are very efficient. It also blew any ideas or assumptions that we had that there were efficiencies to be gained by intercepting goods coming out of Eastern Canada, which we had thought for a number of years would be the case since Eastern Canada is the source of a majority of our goods. And it showed us that we actually could save some money by trading a larger Alberta facility for continued expansion to our Richmond location.

Q: Do you plan to redo this type of optimization on a regular basis?
Mahlman: Yes. We proved to our executives that this was an incredibly valuable exercise, especially bringing it down to the level of detail that we did. Now that we have the model built, it will be a lot less work to do it again. We will need to validate the assumptions and ideas that have come out of the optimization study by rerunning it against the actual numbers. We don't have a defined time line, but probably every couple of years we will look at it again to make sure we are still hitting our plans and objectives. As you well know in this business, the moment you think you have figured out what your profile will look like for the next couple of years, that is exactly when it changes.

Q: Did you learn any lessons during this project that you would like to share?
Mahlman: A couple of things. The master data that you are colleting is probably never as clean as you would like or in the exact format you would like-that's true of any sort of project that touches on IT. So it is always worthwhile to make sure you are tending to your data in an ongoing way. Then it is not such an issue when you do have to recall long history. The second thing, which has always been an issue for London Drugs because of our private nature and our long history, is that we don't trust our partners very easily. The ones that we do pick, after quite a bit of due diligence, we tend to spend a long time with. Most consulting companies have similar competencies and all have good people, but it is whether or not the corporate philosophy and the people fit. We spent a lot of time verifying that ESYNC was the right type of company, the right type of partner for us, and we are extremely excited that our due diligence paid off. We have really been happy with the partnership. So the lesson learned is to not be romanced just in the outcomes that a company can deliver, but to make sure that it matches the culture of your company and that you have similar philosophies. That will save a lot of time.