Optimizing Container Loads For Assorted Products

1. Group products with similar packaging, by way of cartons/boxes on a pallet together. This facilitates some flexibility in mixing.

2. Get historical demand for different SKUs and divide it by the number of pallets required.

3. Adjust pallets ordered at a time to fill a container load and have the same frequency of order a year

Online Ordering for Grocery Pick-Up Service At Physical Stores May Be Better

In a dense cities like Singapore, Chicago and New York where supermarkets are close by, online ordering for pick ups at supermarkets on the way home is the way to go.

Consumers have to get home, and pass their usual supermarket anyway, and online grocers can save on transportation cost and the uncertainty of whether anyone is at home or not. 

What's more, supermarkets can cross sell, upsell and bundle products more easily this way, and consumers are familiar with the quality and freshness of established retail chains.

Read more at http://m.washingtonpost.com/business/economy/value-added-grocery-store-pickups-could-be-next-business-disruption/2013/10/06/bc86a856-2aa4-11e3-8ade-a1f23cda135e_story.html

Operational Opportunities With Advance Export Declaration (AED)

With Advance Export Declaration (AED) requirement, there are many uses for data collected in advance.

1. one can use the advance declaration information to decide how much manpower to allocate as the type of processing required is now known before-hand. Of course, the assumption here is that the time taken to do certain jobs is known.

2. An extension of the above idea is to split services to different counters so that both counters are balanced. This means maximum jobs processed in minimum possible time and resources.

3. Schedules can then be given to freight forwarders to send in their cargo with the minimal wait time.

4. In addition, loads on comtainers/ airline unit loading devices (ULD) can be pretty much optimized in advance, which can coincide with send in schedules. Space is saved this way and a higher throughput is achieved.

5. An extension of this is of course the ability for freight forwarders and shippers to obtain dimension and weight measurements in advance to flow through opportunities, increasing the overall supply chain velocity 

Click Here to Buy A Tomohopper today

Assortment quantity decision to fill containers

For many small and medium sized companies looking to buy associated products from the same supplier to fill up shipping containers, many still use gut feel rotate among products to stock up just to fill the container. This gut feel may or may not take into account variability of past sales or sales forecast ahead.

A scientific approach to apportioning how much extra stock to carry will be using the safety stock factor approach. Where we say the proportion of extra stock to buy for one SKU relative to another SKU is in a ratio of z*stdev of demand for one SKU / z*stdev of demand for another SKU. This is suitable for SKUs that are already in the market for some time and has reliable historical sales numbers. The multiples of SKU purchase would of course, we governed by MOQs and/or order multiples.

If it is a new product, then we should look at it from a cost of overstocking vs cost of understocking. It's ratio will give it the proportion beyond or under demand.

If the SKU is a combination of new and old products, then we have to combine the ratios of the different SKUs together. It is possible because the underlying normal assumption probabilities is the same.

Delivery Scheduling in City

Modified a concurrent scheduler approach to consider different priorities of tasks, and a practical counter to rotate use of resources for continuous performance analytics to minimize resources. 

Concurrent scheduler approach makes more sense compared to the usual Operations Research approach to scheduling/routing, as this approach is more dynamic, takes conservative traffic prediction into account but is responsive to current traffic condition, and promotes continuous improvement.

What is most important, is it is easy for a user to understand to improve continuously.

Damage and Cleaning Image Recognition for Containers

Just finished a course on Machine Learning on Coursera and found that we can train a machine to recognize letters. Much like how Optical Character Recognition is developed. By the same approach, one can do the same, and have already done, for face recognition for images. With some many containers passing through Singapore, and with so many pictures taken of damaged containers for repairs, we can actually use this to train a machine to recognize damaged containers for areas of repair and its cost. One can certainly use current CCTVs or other types of video cameras to capture photos. 

The manpower and processing needed currently to do this will be saved.

Improve Forecast To Lower Safety Stock

We have all been taught the safety stock formula. It is however, not very responsive to use average sales and historical sales to calculate the sigma value. This is because the sales changes with time, and we are all unsure of the period of historical sales to average it over. Additionally, new product launches, promotions and variants all affect actual sales numbers. 

Since the sigma value for safety stock is a measure of the difference between projected vs actual sales, it makes a lot of sense that sigma be derived from forecast vs actual sales. This is a more responsive and forward looking approach as forecast incorporates up-to-date factors like sales, promotions or events. Plus, we are using the actual sales rather than average sales to derive the sigma for safety stock. This has significant implication for re-order point calculation subsequently.

What's more, measuring forecast and actual sales also facilitates Sales & Operations Planning in an organization.

Piece Picking Automation For Online Sales

Piece picking automation is critically important to grow online businesses. Test-bedding on automation equipment and measuring productivity and cost savings assessment needs to be done continuously to cater to changing customer requirements on consumer goods.

Improving Fill Rate with Customer Service Level

We know fill rates and customer service level measures different things. The former measures how much of total orders is fulfilled, while the latter tells you how often, say out of 10 stock replenishments, there will be an out of stock situation for one stock cycle (90% service level). Though they are different, both measures are actually related.

To find the relationship, check if a fill rate is due to a particular order with many lines not filled? or an accumulation of one or two lines not filled over many orders? If it is the former, then we need a much higher service level, if it is a the latter, theyoshinoya 90% will help increase Line Fill rate accordingly.

This line fill rate will have a snow-ball effect to revenue fill rate, and on-time fill rates.

Space-Filling Curve For Delivery Clustering

Knowing that space filling curve connects points in the shortest possible route, the only limitation for companies to make use of it is specific delivery time windows and truck capacity.

One could use their highest capacity trucks to tackle high volume, time sensitive deliveries first as fixed point on a space filling curve, and move on from there.

Order Fulfilment Alignment With Optimized Delivery Routes

I am thinking that all orders, or customers to be specific,  should be link to optimized delivery routes in the first place. So that picking can be according to first-in-last-out sequence and it can be loaded into truck immediately, reducing double handling, space and mistakes.

To realize these possibilities, it is important then map a delivery route through all the possible customers (e.g all possible postal codes). To find the optimal route sequence. This sequence can be indexed and associated with postal codes in any accounting/ERP software.

Too many postal codes? One way to reduce the postal codes is to obtain the distances between postal codes. If it is close enough, give it the same index.

Order Taking Using Answering Machines

Orders coming fast and furious by telephone calls during peak hours? Customer service officers taking down orders on paper and re-keying into your accounts system? Re-keying data wastes time and is mistake-prone?

One good way of doing it right, the first time is to use a telephone answering system to take orders. In this way, your answering system becomes your 24 hours customer service officer to take down orders. It records what your customers want for checking if any disputes arise. One can also listen to it repeating to clarify orders before keying into system.

Staytime Record For Home Service Visibility and Scheduling

In cities like Singapore where services are required in dense, built-up areas, GPS tracking may not be good enough. There could be visits to different units on the same block that will not be picked up using GPS.

GPS tracking is know to give location visibility, and by extension, how long one stays at a location. Duration of stay at a location, or what I call staytime, is very important. It allows planner to schedule work better to maximize resources.

How do we get staytime data for dense, city services. Well, one say it to have arrival and departure forms to fill. This will allow staytimes, and the associated latitude and longitude location to be captured for useful analysis.

Inventory Reduction Opportunities with MOQ and INCO terms

With so many SKUs to manage, and some with high minimum order quantities (MOQ), it is small wonder that companies will buy and hold inventory in bulk, especially when there could be full container load savings.

However, if the Sales Terms, or International Commercial (INCO) Terms agreed is such that the supplier/principal pays for freight, then there could be opportunity to buy more often, and less at a time, to reduce inventory and space requirement.

How do you know? Just plot a graph showing high MOQ/Average demand during leadtimes on the left, and the order frequency in the right for products with CNF and beyond. Do a quick sanity check to compare if high MOQ/Average demand during leadtime translates to lower number of orders (unless it is planned orders). Those products that you see that is ordered infrequently are opportunities and savings.

Forecasting To Reduce Inventory

Setting higher reorder points because you do not know how sales will be like, and you feel it is better to buffer with higher inventory?

This is a common refrain by companies who are not forecasting well. This is especially useful for products with short lead times, and where space is expensive.

The change in sigma in the safety stock allows you to compare better forecast with current safety stock for savings. There are a number of forecast accuracy measures to choose from, ranging from mean average deviation (MAD) to root mean square error (RMSE).

In terms of savings, there will be capital cost, space and it's corresponding operations savings. More accurate forecast means less rework too.