May 5, 2009
BrainTrust Query: Why is computer-based ordering suddenly an in-demand technology application?
By Kevin Stadler,
Vice President of Sales and Marketing, SAF USA, Inc.
Due to the state of the
economy, retailers are scrutinizing technology solutions more closely this
year to determine which applications will provide the best benefits for
their stores. According to the 2009 Supermarket News 15th Annual
State of the Industry Report on Supermarket Technology, food retailers
ranked computer-based ordering first for which new applications they will
either test or launch in 2009, tying with trade promotion management at
16 percent.
According to the report,
computer-based ordering also ranked first in 2008, at 17 percent, for new
applications that were either tested or launched. Hannaford was cited
as a retailer that successfully rolled out a computer-based ordering system
across its chain last year. To date, the retailer has reduced out-of-stocks
up to 70 percent in many of its stores.
This trend is interesting
because the concept of computer-based ordering and its pioneering technology
has been around for more than 20 years. Since reducing out-of-stocks is
an operational imperative, one would expect computer-based ordering solutions
to be deployed universally by now. Instead, only a few retailers have crossed
the chasm from manual to automated store ordering. From what we’ve seen,
the main reason that retailers still remember early computer-based ordering
failures, while others determined long ago that the effort was too great.
However, retailers in
addition to Hannaford that have taken the leap and successfully rolled
out computer-based ordering systems have reported rewards including 60
to 80 percent reductions in out-of-stocks and 25 to 40 percent reductions
in store inventory. They have also been able to process 98 percent of their
order lines automatically.
Discussion Questions:
Do you see computer-based ordering being broadly adopted among retailers
at this point? Is it ready for “prime time”? How would you
weigh the technology’s benefits and weaknesses?
BrainTrust
Recent Discussions
Discussion Questions
Poll
At a minimum, we have seen big interest in replenishment forecasting and store-level forecasting, a prerequisite for successful computer-based ordering. And hearing retailers talk not just about the idea of doing it, but how they’ve done it.
This is just the other side of the technology pendulum’s swing. First we saw a move to centralized buying, to help consolidate and introduce efficiencies on the supply side. But we gained it at the loss of the idiosyncrasies of each individual location. That’s half of what this most recent push into customer-centricity has been all about: regaining that personal touch in stores. The only way to do that is through technology, or you’ll have to give back the efficiencies you gained in central buying.
But this is really round two of CAO/CBO. The first round was about two years ago, when retailers realized that while they want it, they needed to have much cleaner data sources to make it work. The last two years have generally seen a lot of effort around cleaning that up, so it’s not surprising to me that store replenishment is back on the agenda once more.
A major reason that many retailers have resisted it is that they do not have good enough scan accuracy. While scanning has improved, has it improved enough to drive replenishment? If not, it will quickly show up if the scan data is driving orders and inventory replenishment. While it may not make any financial difference if different flavors or varieties at the same price are scanned in mass, but it makes a big difference when it comes to ordering product. Scanning needs to continue to be more of a science than an art.
It’s interesting that today’s discussions include this one on computerized ordering and the other on Kroger’s SKU rationalization effort. Store replenishment is driven by shelf allocation. There are two types of items; ones whose minimum order quantity (case pack) exceeds delivery interval sales and ones whose case pack does not exceed the sales between delivery (so more than one case must be ordered). When it comes to shelf allocation there are three types of items; ones whose shelf allocation exceeds the sales between deliveries and will not require stocking in between deliveries, ones whose shelf allocation is less that inter-delivery sales but sufficient to reach next stocking period (e.g. night crew) so they will they will be restocked from backroom stock during normal stocking, and finally the sale or very fast moving items such as dairy and produce items which will require restocking between normal stocking activity.
Over 90 percent of the items in a supermarket will sell less than a case a week, meaning their minimum order quantity exceeds sales across multiple scheduled deliveries. For those items, no forecasting is necessary and a simple re-order point (KanBan) methodology is both adequate and simple to implement.
For the remaining 10 percent of the items, a flexible approach which recognizes seasonal a promotional effects on the population of items which fall into this category is necessary. Forecasts are necessary to predict how many cases should be ordered. But it is also necessary that the items in the third shelf allocation category be watched closely. Stores need to implement procedures to recognize when items without enough shelf allocation to reach the next normal stocking schedule need stocking. This can be as simple as a list of the items on sale each week and a process of walking the store periodically to check their shelf inventory. Study after study have shown that these are the items generating most of the out of stocks.
This is not rocket science, we are selling groceries.
This is excellent news and I truly hope the trend continues as the store is both the beginning and the end of the Extended Retail Supply Chain. Information flow starts at the store and product delivery ends there. That being said, the store is just the first step towards total supply chain excellence. The key is to link the store CAO system to its supply locations (retail DCs, manufacturers DCs & factories) inside a single planning system to really get a bang for the buck.
Computer based store ordering is not new, neither are the problems. Computer Assisted Ordering (CAI) has been in operation for years, but needed an order writer to adjust just about every order. The reasons are well known.
Retailers don’t have an accurate store item inventory. Scanning is not accurate due to the multiple key and other checkout errors. Damage & Unsalable don’t get recorded in a timely manner. Shrink goes undetected until there is a physical inventory. No one wants to expend the labor to inventory the backroom. Last but not least is many retailers storing movement information at the weekly, not daily level in their computer systems. Couple all problems with a fixed store delivery schedule and these problems only compound into an unworkable process.
For computer-based store ordering to really work requires a history of daily movement by item. Change the order writer into an inventory manager performing cycle counts to keep the inventory accurate. Eliminate the store fixed delivery schedule and replace store replenishment with need. When a store sells a truckload, send a truckload. Not only will Out-Of-Stocks be reduced, but backroom inventory should be greatly reduced. This assumes shelf allocation has some relationship to item movement.
One critical issue in Computer Assisted Ordering is Phantom Inventory, when the system believes there is product available on the shelf when there is actually a stock-out. Stocking crews can easily face-over to hole and the stock-out can persist for days or weeks. The slowed sales may even drive discontinuation.
Solving this issue requires three things:
1) Computer Assisted Demand Modeling. Tools now exist to model daily shopper demand at a store/day/item level based on merchandising causals (TPR, display, feature) and other environmental causals (weather, sporting event, holiday). This added intelligence provides the CAO solution ‘front-end’ demand that can be converted to store orders. As the shopper demand is recalculated daily for the next 2-4 weeks, any forecast errors are quickly detected.
2) A practical approach to space allocation. The shelf holding inventory should equal (the minimum replenishment quantity (a case)) + (best selling day volume X replenishment cycle days) + (potential safety stock). This approach reduces store labor as full cases go directly to the shelf, significantly reduces back-room inventory (a root cause for phantom inventory – the CAO system shows inventory that is not on the shelf), and improves in-stock conditions.
3) A fresh look at promotion inventory flows. Most promotions can be serviced from their home inventory locations, particularly if the CAO system is aware of the incremental demand. For those items that required additional floor inventory to support consumer demand, initial force-out quantities should be minimal, followed by pulsed replenishment. In our experience, this both serves the merchandising needs and results in minimal post-promotion excess inventory–inventory that can get lost in the back room, but the CAO system shows as inventory.
The result: Better in-stock, faster turns, happy shoppers.
After reading all the reasons why CAO is not working, I cannot help but ask the following question. “If the largest retailer in the world is doing it and doing it quite well, (it’s not perfect but it’s working) why then are other retailers holding back?”
Could it be a question of understanding its importance to the overall management of the extended retail supply chain? When there is a will, there is a way.
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?
At a minimum, we have seen big interest in replenishment forecasting and store-level forecasting, a prerequisite for successful computer-based ordering. And hearing retailers talk not just about the idea of doing it, but how they’ve done it.
This is just the other side of the technology pendulum’s swing. First we saw a move to centralized buying, to help consolidate and introduce efficiencies on the supply side. But we gained it at the loss of the idiosyncrasies of each individual location. That’s half of what this most recent push into customer-centricity has been all about: regaining that personal touch in stores. The only way to do that is through technology, or you’ll have to give back the efficiencies you gained in central buying.
But this is really round two of CAO/CBO. The first round was about two years ago, when retailers realized that while they want it, they needed to have much cleaner data sources to make it work. The last two years have generally seen a lot of effort around cleaning that up, so it’s not surprising to me that store replenishment is back on the agenda once more.
A major reason that many retailers have resisted it is that they do not have good enough scan accuracy. While scanning has improved, has it improved enough to drive replenishment? If not, it will quickly show up if the scan data is driving orders and inventory replenishment. While it may not make any financial difference if different flavors or varieties at the same price are scanned in mass, but it makes a big difference when it comes to ordering product. Scanning needs to continue to be more of a science than an art.
It’s interesting that today’s discussions include this one on computerized ordering and the other on Kroger’s SKU rationalization effort. Store replenishment is driven by shelf allocation. There are two types of items; ones whose minimum order quantity (case pack) exceeds delivery interval sales and ones whose case pack does not exceed the sales between delivery (so more than one case must be ordered). When it comes to shelf allocation there are three types of items; ones whose shelf allocation exceeds the sales between deliveries and will not require stocking in between deliveries, ones whose shelf allocation is less that inter-delivery sales but sufficient to reach next stocking period (e.g. night crew) so they will they will be restocked from backroom stock during normal stocking, and finally the sale or very fast moving items such as dairy and produce items which will require restocking between normal stocking activity.
Over 90 percent of the items in a supermarket will sell less than a case a week, meaning their minimum order quantity exceeds sales across multiple scheduled deliveries. For those items, no forecasting is necessary and a simple re-order point (KanBan) methodology is both adequate and simple to implement.
For the remaining 10 percent of the items, a flexible approach which recognizes seasonal a promotional effects on the population of items which fall into this category is necessary. Forecasts are necessary to predict how many cases should be ordered. But it is also necessary that the items in the third shelf allocation category be watched closely. Stores need to implement procedures to recognize when items without enough shelf allocation to reach the next normal stocking schedule need stocking. This can be as simple as a list of the items on sale each week and a process of walking the store periodically to check their shelf inventory. Study after study have shown that these are the items generating most of the out of stocks.
This is not rocket science, we are selling groceries.
This is excellent news and I truly hope the trend continues as the store is both the beginning and the end of the Extended Retail Supply Chain. Information flow starts at the store and product delivery ends there. That being said, the store is just the first step towards total supply chain excellence. The key is to link the store CAO system to its supply locations (retail DCs, manufacturers DCs & factories) inside a single planning system to really get a bang for the buck.
Computer based store ordering is not new, neither are the problems. Computer Assisted Ordering (CAI) has been in operation for years, but needed an order writer to adjust just about every order. The reasons are well known.
Retailers don’t have an accurate store item inventory. Scanning is not accurate due to the multiple key and other checkout errors. Damage & Unsalable don’t get recorded in a timely manner. Shrink goes undetected until there is a physical inventory. No one wants to expend the labor to inventory the backroom. Last but not least is many retailers storing movement information at the weekly, not daily level in their computer systems. Couple all problems with a fixed store delivery schedule and these problems only compound into an unworkable process.
For computer-based store ordering to really work requires a history of daily movement by item. Change the order writer into an inventory manager performing cycle counts to keep the inventory accurate. Eliminate the store fixed delivery schedule and replace store replenishment with need. When a store sells a truckload, send a truckload. Not only will Out-Of-Stocks be reduced, but backroom inventory should be greatly reduced. This assumes shelf allocation has some relationship to item movement.
One critical issue in Computer Assisted Ordering is Phantom Inventory, when the system believes there is product available on the shelf when there is actually a stock-out. Stocking crews can easily face-over to hole and the stock-out can persist for days or weeks. The slowed sales may even drive discontinuation.
Solving this issue requires three things:
1) Computer Assisted Demand Modeling. Tools now exist to model daily shopper demand at a store/day/item level based on merchandising causals (TPR, display, feature) and other environmental causals (weather, sporting event, holiday). This added intelligence provides the CAO solution ‘front-end’ demand that can be converted to store orders. As the shopper demand is recalculated daily for the next 2-4 weeks, any forecast errors are quickly detected.
2) A practical approach to space allocation. The shelf holding inventory should equal (the minimum replenishment quantity (a case)) + (best selling day volume X replenishment cycle days) + (potential safety stock). This approach reduces store labor as full cases go directly to the shelf, significantly reduces back-room inventory (a root cause for phantom inventory – the CAO system shows inventory that is not on the shelf), and improves in-stock conditions.
3) A fresh look at promotion inventory flows. Most promotions can be serviced from their home inventory locations, particularly if the CAO system is aware of the incremental demand. For those items that required additional floor inventory to support consumer demand, initial force-out quantities should be minimal, followed by pulsed replenishment. In our experience, this both serves the merchandising needs and results in minimal post-promotion excess inventory–inventory that can get lost in the back room, but the CAO system shows as inventory.
The result: Better in-stock, faster turns, happy shoppers.
After reading all the reasons why CAO is not working, I cannot help but ask the following question. “If the largest retailer in the world is doing it and doing it quite well, (it’s not perfect but it’s working) why then are other retailers holding back?”
Could it be a question of understanding its importance to the overall management of the extended retail supply chain? When there is a will, there is a way.