Over the years, working for and with numerous manufacturing companies, I’ve seen many supply chain practices that cost companies money. Over the next several weeks, I’ll outline these issues and discuss some ideas around how to avoid these practices. You can find the previous posts here:
- Reason #1: Offshoring without getting the full picture
- Reason #2 Poorly executed or non-existent sales and operations planning
- Reason #3 Not having end-to-end supply chain visibility
- Reason #4 Making key decisions by modelling the supply chain in Excel
- Reason #5 Not having a supply chain risk management process
Reason #6 Not effectively managing inventory
I had to throw out some carrots yesterday. I hate throwing food out but there was nothing to be done for it…all I can say is that I’m glad the carrots were in a bag….and it didn’t leak. That got me thinking about why I was throwing away what had been perfectly good food;
- I had forecasted needing a certain amount, but the customers (my family) didn’t take what I’d forecasted.
- I thought we would want carrots, but everyone wanted broccoli…which I didn’t have.
- I lost track of how many carrots we had and ended up buying more when we really didn’t need any.
- Spoilage can happen. In the case of my carrots, there was a limited shelf life – but they could have been dropped or stolen (hey, it could happen!).
That was carrots. All in all, it cost me a couple of dollars. Unfortunately, all the same kinds of things can happen to your supply chain inventory. Except that your inventory costs millions of dollars. Those of you that manage inventories know how hard it can be to get the quantities just right. If you maintain too little inventory, you have stockouts, line stoppages and unhappy customers. If you have excess inventory, it ties up working capital and is at risk of damage and obsolescence. The worst possible world is when you have too much of something you don’t need, and too little of something you do need. So what strategies are out there to maintain inventories at the “right” level? There are many but let’s focus on some of the high runners;
- Sales and Operations Planning – How many times have you seen this scenario play out? Marketing sees an opportunity and plans a huge promotion for Product B. Operations is going full out building anticipation inventory for Product A because Product A’s demand always goes up this time of year. By the time operations realizes that marketing is promoting a different product, they already have too much inventory of Product A and don’t have enough time to make enough Product B to satisfy the demand driven by the promotion. If this company had an effective S&OP process, operations and marketing would have been aligned as soon as Marketing had approved the promotion, and would have had the right amount of inventory of the right product.
- Better forecasts – Forecasts are always wrong! True. But sometimes they can be less wrong…and the more accurate your forecast, the more likely it is that you’ll be building the right quantity of the right products at the right time. Forecasting is hard, however, advanced tools like statistical forecasting algorithms, collaborative forecasting tools and forecast accuracy measures and what-if scenarios helps guide demand planners to a more accurate set of numbers.
- Lead time reductions - Supply chain improvements can actually help improve forecasting! Well,actually it would reduce the impact of bad forecasting when you are a make to stock shop. How does that work you ask? Imagine you were asked to accurately predict the weather for this time next year. Pretty tough right? What about six months from now….still hard. What about next week? Getting easier. How about tomorrow? No problem (usually)! In a make to stock environment, if I have a 6 month cumulative lead time, my forecast is being used to buy inventory today for something I’m going to sell in 6 months. If through process improvements, I can reduce my lead time to 2 months, my accuracy will be much better where it really matters; during my cumulative lead time.
- Better/faster planning - While there are things you can (and should) do to improve your forecasts, you are never going to realize truly accurate forecasts. For example, a surveyfrom 2012 showed that average forecast error by industry ranged from 15% for retail to 39% for manufacturing/industrial and consumer packaged goods. Forecast error for most other industries was around 30%.One of the problems with poor forecast accuracy is that today’s legacy systems are unable to respond fast enough to satisfy demand that is in excess of forecast. This leads to a) higher than necessary inventory levels as we maintain higher inventories on those items with the highest variability and forecast error or b) lower than acceptable customer service. Neither are good results.
So how do we respond faster? There are multiple capabilities your demand system must have to allow faster response to demand fluctuations;
- Visibility across the enterprise – to be able to respond effectively, your planning system must contain all data across all plants, regardless of the source system. Responding quickly means knowing what you have and what you don’t have. If you have to wait hours or days to get a report off a remote system, you can’t respond.
- Always on analytics – Imagine creating an excel model but every time you made a change, you had to wait 6 hours to see the impact. It wouldn’t be very useful, right? Yet this is what we accept from our ERP systems every day. To simulate effectively, you need to be able to see the result of a change as soon as that change is made. Not only must the calculations be fast (seconds not hours) but the calculations must be configurable enough to allow you to model ERP results from any ERP system (what’s the point of figuring out what to do, if you can’t replicate the results in your execution system)
- What-if scenarios including scenario comparison – There is never only one answer to complex problems like supply change. Being able to try out multiple approaches very quickly and compare these approaches means that you can quickly zero in on the best answer.
- Collaboration – No one person has the knowledge of the entire supply chain in their head. You must be able to rely on others to help figure things out. You must be able to determine who needs to be involved, then share the appropriate scenarios and information with those people if you want to respond quickly (and confidently).
- Alerting based on impact, not on the event – There are a lot of things vying for our attention today. So many, in fact, that we don’t have time to deal with items that aren’t truly important. Traditional ERP systems drown us in frivolous messages; this supply order is 1 day late, this customer added an order, this job finished on time, etc, etc. This is not important information –and as a rule can be relegated to summarized reports. What is critical is: what the impact of these events? For example, if that order is one day late, it impacts $3 million in customer orders. That’s what you want to know. If the order is replenishing safety stock – who cares?
- Inventory planning and optimization – Safety stock traditionally has been a pain to calculate – as a result many people didn’t. They either set the safety stock level once – and forgot about it or did a best guess at what the Safety Stock should be. Inventory Planning is a relatively new area where safety stock is statistically determined based desired customer service levels and on supply and/or demand history. Traditionally, Safety stock was calculated a single level at a time and didn’t consider the stock of the parent or component item when calculating its own stock level. Multi-Echelon optimization looks at the inventory for a family of parts and determines where it makes the most sense to locate inventory for individual items within that family and potentially lowers the overall inventory for the family.
- Inventory accuracy – similar to my carrot analogy - we all have had situations where you go to the store, buy some goods – then discover that you have 6 cans of the thing you just bought hiding behind the peanut butter. Or worse, you THINK you have 6 cans of thing you need for supper and you don’t pick more of it up – then you discover that someone (maybe you?) ate it and you actually have none. In supply chain, the same thing happens – but the cause is inaccurate inventory records and the cost can be huge. How do inventory records get out of alignment? In a previous life, I used to work with the operations team and track down inventory records. The biggest culprit was human error; incorrect quantities, incorrect BOMs, spillage, waste, etc.There are two approaches to maintaining accurate records;1. Annual physical inventory – This is a traditional inventory management technique where you take several days, tag all of the items in inventory and have some poor guy count the items, write the count on the tag and turn the tag into a central team that updates the records. If there is a problem, the poor guy may be asked to go out and count the items again. There are some problems with this approach;
- While the inventory is being done, the factory cannot run. This means inventory must be done over the weekend or the factory needs to be shut down.
- Physical inventories are not fun. It’s tedious, boring, dirty, nasty work (speaking as someone who’s done it). It’s often performed by people not necessarily tied to the inventory function. It’s difficult to be precise counting thousands of different parts in the course of a few days. It’s very likely that a significant number of the counts will be wrong.
- The root cause of the error (why the inventory is wrong) is seldom ever caught and as such, doesn’t get corrected.
- The supply chain continues to function while the cycle count is done
- The count is performed by inventory specialists that know the inventory, are used to counting and are incented to get it right.
- Key parts are counted more frequently and therefore will be more accurate.
- When a discrepancy is found, the team seeks to understand why the error occurred and ideally determines what changes they need to make to prevent the error from happening again.