July 17, 2008
Supply Chain Digest: Order and SKU Velocity Profiling is Key to Picking System Design and Technology
By SCDigest Editorial Staff
Through a special arrangement, what follows is an excerpt of a current article from Supply Chain Digest, presented here for discussion.
There are a wide range of potential picking system options, from manual, paper-based systems to highly automated approaches. So how do you select which option(s) are right for you? For many companies, the question is: where do you begin? The smart answer for most companies: SKU and order activity profiling.
The process of gathering and analyzing data is about both order profiles (orders per day/shift, lines per order, items per line, etc.) and individual SKU activity (volumes, breakdown by unit of measure, etc.).
Let’s state right up front that although order and SKU profiling is key to order picking system design and technology selection, few companies seem to have the data readily available, even in this age of ERP, WMS, data warehouses and other software systems, which one would think would contain the required information. In some cases, the data is available or largely available, but is in different software systems. The data from these different sources must be merged. In other cases, key data is not available or is very hard to get at.
In all cases, companies must carefully look at both what data is available and how accurate that data is.
According to Cliff Holste, SCDigest’s Material Handling System Editor, to determine order picking strategies and technologies, companies need to obtain and analyze the following types of data:
- Order mix distribution (family mix, handling unit, order increment)
- Lines per
order distribution - Cube per order distribution
- Lines and cube per order distribution
Analysis should, in part, develop a SKU velocity profile, similar to the chart below. The analysis will then also typically involve breaking this data down into movement by different handling/ picking units, such as full pallet, full case and split case volumes.
The next step is an iterative process, and is constrained by many factors, especially if the analysis is for an existing building rather than a “green field site.” It may also be that a company wants to look only at one area, such as piece picking. A comprehensive analysis will include both analyzing at where some group of products (such as the Very Fast full case movers) will be stored, in what type of storage mode (e.g., double deep pallet flow rack, half pallet, etc.) and what picking strategy/technology will be used (voice or RF pick to truck, pick by label, voice or RF pick-to-belt, pick-to-light, etc.).
The analysis can get very sophisticated; for example, in some businesses it makes sense to look at items that are almost always ordered as single line items, and store those in a separate area of the DC. It can also pay to look at SKU/Order relationships – a relatively slow moving SKU might make sense to store in the high-velocity area if when it is ordered it almost always is ordered with a Category A product.
Discussion Question: Is SKU and order profiling analysis the key to effective order picking design? What are some smart practices for doing it well?
BrainTrust
Recent Discussions
Discussion Questions
Poll
Many warehouses spend big bucks on technology but won’t consider gainsharing with their staff. In other words, millions of dollars for equipment, software, and real estate, but no bonuses for outstanding human achievement. Great productivity can often result from hourly, daily, and weekly productivity measurement and incentive policies. If the pickers know they can make X% more by picking Y% more per hour, great savings can be made.
Useful thought process here, I think. But we should remember that turn rates originate in the stores, and reflect consumer demand. The supply chain serves the stores, not the other way around.
Differentiating between faster and slower moving items is material to the shelf replenishment process too. Slow movers tend to accumulate shelf inventory in excess of demand, while the fastest movers run chronically out of stock. Both conditions coexist on the same shelves and within the same supply chain.
This leads the discussion back to the In-Store Implementation issue I’ve been hammering at so much recently. Exclusive focus on the distribution and logistics portions of this challenge will yield unsatisfactory outcomes. Integrating a plan-do-measure discipline for shelf-level compliance will lead retailers to sounder practices, both in the stores and up the supply chain.
Profiling and SKU/order analysis plays a huge part in order picking, especially piece picking (picking one or two items at a time) and less important in full case picking. There are other elements that are also important and play a critical part in designing efficient and effective warehouse designs. In retail, a huge item is seasonality. Something that moves fast two months out of the year, like an ice scraper may not move at all the rest of the year.
Another factor is customer service commitments. If the time between order placement is short, a better design becomes more critical vs. a long lead way. When identifying the best order picking methodology, technology and storage medium many factors weigh in to give you the optimum design and certainly SKU/order analysis is a part of the total equation.
Of course….
The analysis of item movement and order sizes also has to be repeated for major seasonal variations. The layout of items used in the summer is not efficient for the winter.
But there is another very subtle purpose for doing this type of analysis that can be used to improve both warehouse efficiency and reduce transportation costs. The goal of the warehouse is to minimize selector travel time, the goal of transportation is to minimize empty truck space. Both goals can be better met by management of the ordering process. The extreme example is preventing stores from ordering potatoes on weekend deliveries. There is no reason why long shelf life produce items should take up space on limited weekend deliveries. Likewise, by properly managing the days stores can order shelf stable items, the retailer can increase the “pick density” and truck utilization for inbound loads. In effect, restricted ordering creates a “virtual slow mover warehouse” by concentrating orders for low priority items on specific days and loads.
The challenge here is that the benefit at the warehouse comes at an inconvenience for the store. In a wholesaler supported environment there has to be an incentive for stores to accept the inconvenience. This can be implemented with a surcharge on items ordered on their “day off.” Then the retailer has an incentive to order the slow movers on the designated days.
First, a quick historical note. The first grocery automatic picking system was the SI Ordermatic. You computer punched cards then loaded the cards into the system computer. The system picked cases and sent them on a belt for card or pallet loading. The SI Ordermatic was removed from grocery distribution centers due to the rails being welded and the product mix change over time that made it impossible to use the machine. So product mix alone can change the picking needs even before an increase in SKUs.
The point missed by most is when the SKU count increases, the velocity decrease for other items. The mistake most distribution centers make is not performing an analysis and reset–at a minimum of–quarterly. Then the building adjusts with small changes, not a major re-set. Although I have preached this for years, I have yet to see a distribution center actually do this. Only after years of stuffing SKUs anywhere they can and the pain is great, do they consider a re-set.
The picking method issues are simple. Picking labor is 50% or more of distribution center labor. Therefore, labor savings have the highest priority. Second is accuracy of pick. Picking and sending the wrong item to a store does not help the store and now you have two items in the distribution center with inventory errors.
The Return on Investment approach is best for evaluating picking methods. The greater the labor saving, the easier it is to justify investment. The slow selling items will rarely justify any investment.
I would just add that many/most understand the theory and benefit of SKU/Order profiling–but few actually muster up the energy and resources to do it effectively.
Most get the basics–e.g., highly seasonal items–correct, but the real advantages come from getting granular.
I have never yet seen a company that did this for the first time in a long while that didn’t learn some things that challenged their conventional beliefs about their business and distribution patterns.
There is clearly a mismatch between the level of benefit (high) and consistent practice of the art and science (low). With some exceptions (Canadian Tire in the retail world is a great example, as is O’Reilly Auto Parts, among others) it takes a major event (e.g., new distribution center) for companies to do this profiling well.
More Discussions
As Boomers ‘Age Out of Peak Purchasing,’ How Can National Brands Retain Interest With Gen Z?
According to a recent First Insight study titled “Is Gen Z Still Choosing Your Brand,” while…Is Supply Chain Volatility Here To Stay?
Can we expect an ongoing volatility to remain in terms of retail supply chain?Has Same-Day Delivery Become an Expectation Across Retail?
Are expectations around retail delivery speed in flux?
Many warehouses spend big bucks on technology but won’t consider gainsharing with their staff. In other words, millions of dollars for equipment, software, and real estate, but no bonuses for outstanding human achievement. Great productivity can often result from hourly, daily, and weekly productivity measurement and incentive policies. If the pickers know they can make X% more by picking Y% more per hour, great savings can be made.
Useful thought process here, I think. But we should remember that turn rates originate in the stores, and reflect consumer demand. The supply chain serves the stores, not the other way around.
Differentiating between faster and slower moving items is material to the shelf replenishment process too. Slow movers tend to accumulate shelf inventory in excess of demand, while the fastest movers run chronically out of stock. Both conditions coexist on the same shelves and within the same supply chain.
This leads the discussion back to the In-Store Implementation issue I’ve been hammering at so much recently. Exclusive focus on the distribution and logistics portions of this challenge will yield unsatisfactory outcomes. Integrating a plan-do-measure discipline for shelf-level compliance will lead retailers to sounder practices, both in the stores and up the supply chain.
Profiling and SKU/order analysis plays a huge part in order picking, especially piece picking (picking one or two items at a time) and less important in full case picking. There are other elements that are also important and play a critical part in designing efficient and effective warehouse designs. In retail, a huge item is seasonality. Something that moves fast two months out of the year, like an ice scraper may not move at all the rest of the year.
Another factor is customer service commitments. If the time between order placement is short, a better design becomes more critical vs. a long lead way. When identifying the best order picking methodology, technology and storage medium many factors weigh in to give you the optimum design and certainly SKU/order analysis is a part of the total equation.
Of course….
The analysis of item movement and order sizes also has to be repeated for major seasonal variations. The layout of items used in the summer is not efficient for the winter.
But there is another very subtle purpose for doing this type of analysis that can be used to improve both warehouse efficiency and reduce transportation costs. The goal of the warehouse is to minimize selector travel time, the goal of transportation is to minimize empty truck space. Both goals can be better met by management of the ordering process. The extreme example is preventing stores from ordering potatoes on weekend deliveries. There is no reason why long shelf life produce items should take up space on limited weekend deliveries. Likewise, by properly managing the days stores can order shelf stable items, the retailer can increase the “pick density” and truck utilization for inbound loads. In effect, restricted ordering creates a “virtual slow mover warehouse” by concentrating orders for low priority items on specific days and loads.
The challenge here is that the benefit at the warehouse comes at an inconvenience for the store. In a wholesaler supported environment there has to be an incentive for stores to accept the inconvenience. This can be implemented with a surcharge on items ordered on their “day off.” Then the retailer has an incentive to order the slow movers on the designated days.
First, a quick historical note. The first grocery automatic picking system was the SI Ordermatic. You computer punched cards then loaded the cards into the system computer. The system picked cases and sent them on a belt for card or pallet loading. The SI Ordermatic was removed from grocery distribution centers due to the rails being welded and the product mix change over time that made it impossible to use the machine. So product mix alone can change the picking needs even before an increase in SKUs.
The point missed by most is when the SKU count increases, the velocity decrease for other items. The mistake most distribution centers make is not performing an analysis and reset–at a minimum of–quarterly. Then the building adjusts with small changes, not a major re-set. Although I have preached this for years, I have yet to see a distribution center actually do this. Only after years of stuffing SKUs anywhere they can and the pain is great, do they consider a re-set.
The picking method issues are simple. Picking labor is 50% or more of distribution center labor. Therefore, labor savings have the highest priority. Second is accuracy of pick. Picking and sending the wrong item to a store does not help the store and now you have two items in the distribution center with inventory errors.
The Return on Investment approach is best for evaluating picking methods. The greater the labor saving, the easier it is to justify investment. The slow selling items will rarely justify any investment.
I would just add that many/most understand the theory and benefit of SKU/Order profiling–but few actually muster up the energy and resources to do it effectively.
Most get the basics–e.g., highly seasonal items–correct, but the real advantages come from getting granular.
I have never yet seen a company that did this for the first time in a long while that didn’t learn some things that challenged their conventional beliefs about their business and distribution patterns.
There is clearly a mismatch between the level of benefit (high) and consistent practice of the art and science (low). With some exceptions (Canadian Tire in the retail world is a great example, as is O’Reilly Auto Parts, among others) it takes a major event (e.g., new distribution center) for companies to do this profiling well.