Video

Using Inventory Optimization to Find Hidden Money in the Supply Chain

Video Transcription

Dan:
Hi, this is Dan Gilmore. Supply Chain Digest and the Supply Chain Television Channel here for another very special supply chain thought leadership series of broadcasts. Here part 2 on achieving found money in your supply chain. What a great concept that is. We’re really talking about how you can find hidden money very fast and often very easily in your supply chain through design and modeling techniques. Very pleased to be here again with Toby Brzoznowski. He’s executive vice president at LLamasoft, they’re a leader in supply chain network design software and planning software. Toby as always thanks for joining me today.

Toby:
Thanks for having me.

Dan:
Okay, we had a great discussion in part one on this on the concept of product flow path optimization and how that could provide some kind of easy wins for companies. Part two here today on demand segmentation and inventory right sizing. Before we do that, just real quickly, revisit for us with today how you guys came up with this idea of found money.

Toby:
Sure, sure. The whole idea is modeling your supply chain and taking all of this data and starting to get visibility into what’s there and really for many times focusing on these long term big ideas is where the focus has been. What we saw along the way was many people were getting quick wins. They were finding things they didn’t know existed as opportunities for cost savings in the supply chain. We started to look at what were the trends of those things and categorize them and that’s where we came up with this concept that you know they are a lot of found money opportunities that are very much like putting on that coat after a wedding that you hadn’t worn in six months and pulling out a $20 bill. This idea that it’s fast, it’s easy, you didn’t have to work really hard to get it and it’s unexpected. These same attributes applied to the modeling.

Dan:
Real quickly, last time it wasn’t a $20 bill that one company found it was a $20 million dollar bill, they found by switching their port strategies modestly. Very exciting stuff.

Today we’re talking about demand segmentation, inventory right sizing. What’s that all about?

Toby:
Sure. Segmentation is used a lot, the term segmentation. It can mean a lot of different things depending on what you’re talking about. You’ll hear people talking about segmenting their supply chain. It could be segmenting your customers based on their location, or their attributes, or their different buying behaviors. It could be segmenting your products based on their attributes. There’s small and there’s big and there’s heavy, you have all these different characteristics about products. It might be supplier segmentation based on very specific strategic behaviors that you want to have or relationships that you want to have with customers or again their locations or commodities that they’re providing.

The thing that we’re focusing on very specifically today is demand, buying behavior. Trying to segment your supply chain and look for categories of like buying behaviors and using that then to apply the appropriate inventory strategies to get you your best opportunities for improvement.

Dan:
Just explain that a little bit more, different buying patterns. Give me some examples.

Toby:
We have an applied research team that has spent a lot of time just breaking down demand patterns. What you’ll find is historically there’s been a lot of focus of inventory optimization tools and just assuming everything is normal. The buying behavior is normal but then we’re going to apply all these advanced techniques to solving them. What we saw is in many cases, over 50% of the buying behavior isn’t normal. It’s very different. There might be smooth for certain characteristics but there’s a lot of erratic behavior. There’s things that are very lumpy, I get, I know large quantities but then it goes dormant for a while. Then I get small quantities then I get these unit size behaviors. These different types of behaviors all have basically different types of curves. Those curves need to be treated differently and that’s really the main thing that we’ve been focusing on.

Dan:
It’s really treating service policies, safety stock policies, etc. different and in a more nuance way than just sort of your A, B, C, D type of product classification.

Toby:
Absolutely.

Dan:
Very cool. I think that should resonate with a lot of the viewers out here today. You have some examples where you’ve worked with companies on this kind of problem and found some money for them?

Toby:
Yeah.

Dan:
Give us some examples.

Toby:
Absolutely. Inventory optimization can be used in many industries. Obviously there’s certain types of products where you just can’t store inventory, you’ve got to move it whether it be shelf life issues or what have you. Anybody that for the most part makes or moves things deals with inventory so what we really started to focus on was taking a look at, whether it be retail or manufacturing or consumer goods, or chemical, and starting to first look at the demand behaviors. Then what is the structure of your supply chain. Then what are the possible ways in which you can stock inventory.

This first example is really a manufacturing situation where I have multiple manufacturing locations consolidating into a distribution center and then out to hubs and eventually out to DCs or out to customers. In this situation just by starting and looking at the demand characteristics you start to see some clusters emerge. A cluster’s about these are my fast movers with normal demands, low variability. These are my slow movers that have very erratic behavior. They have very clumpy or unit sized demand. These are the ones that it seems to be very consistent in high volume but it’s extremely erratic. Sometimes it’s unit size, sometimes it’s very large. You break these buying behaviors into categories based on the skew, based on the location, and based on the customer that’s ordering them. You start to come up with clusters of orders.

Once you come up with those clusters of different ordering behaviors, then what you can do is you can apply the appropriate strategy. Whether it be these here that are that are fast movers with high volumes and very regular behavior that are also high margin, perhaps I’m going to stock those closer to my customers. I’m going to make sure I have 99 plus percent service and then I’m going to apply the appropriate inventory optimization algorithms to apply. Where you’re going to find others that have very erratic behavior, very potentially smaller movers that maybe even have lower margin you might be centralizing those. You’re coming up with a totally different strategy. Then what you do is when you come up with the clusters you apply the appropriate policy. It’s not always about min max, sometimes it might be I’m going to have fixed reorder points and fixed reorder quantities. Sometimes I’m going to apply very unique ordering behaviors based on the erratic movement.

What we found in this situation was that the manufacturers was stocking their products in too many locations in most situations. In many cases by pulling it back and postponing products they were finding multiple millions in costs, dollars saved every year.

Dan:
Yeah. I think that’s true and every stocking point you have is going to raise your inventory levels the whole things be equal and things like that.

Toby:
Yeah.

Dan:
I just don’t think enough companies look at it the way you’re doing there and the whole thing, just emphasize here on too, Toby, we’re talking about this found money and the point that adheres me again is you’re not talking about moving distribution centers or building new buildings or whatever. You’re talking about just making some somewhat nuance changes to policies or whatever and somehow a lot of money is showing up.

Toby:
That’s the whole point. It wouldn’t fit the three attributes that we said for found money which is fast, easy, and unexpected. It might be unexpected but it’s certainly not easy and it’s certainly not fast if you have to open five new facilities and then put staff in places and completely redeploy all your inventory. Yeah, what we’re talking about is stock it in different quantities in different locations.

Dan:
Very good. Another example?

Toby:
Yeah, retail is a very common situation where you can leverage inventory optimization technology. In this case here the company had already done some level in their own mind of categorizing their products A, B, C, you know D products. They had done that a lot based on volume for the most part. They weren’t looking so much at buying behaviors. They weren’t looking at it might be high volume but it might be in very erratic type behaviors. They weren’t looking at how promotions in many cases would impact what those volumes are and how and where they needed to stock things. In the end they had about 6 or 7 different service level targets. You know, my A products I have a 99% service target. These A1 products all the way down to an 80% or whatever it might be for your lowest classed items.

What we did was we said all right let’s keep those same categories first of all and let’s just apply better math behind the inventory as to how you’re optimizing. So we kept the same categories, we applied a different algorithm based on the appropriate service level targets. What we found was in, I think it was 7 out of the 8 categories, we were finding situations where their overall stock went down. Their overall, in this case, on hand inventory and working capital went down but it wasn’t across the board. If they had seven DCs that were stocking this product there might be five of them where the stock went down there might have been two that went up quite a bit.

The other thing that we found, was those A products, they were under stocking it. We actually were suggesting stocking a whole lot more than what they had to achieve their targets. The end result was multiple millions of reduce work in capital but it was kind of random or perceivable random as to where those things moved.

Dan:
True.

Toby:
When you get the model that has that skew level detail and the demand behavior. That’s where those things can start to jump out at you. That’s kind of the other message behind this is a modeling technology is what you need to actually uncover those things.

Dan:
Yeah, absolutely, agree with you 100%. Okay you’ve convinced me, this is a very good thing for many companies to do. Inventory optimization is in fact kind of a hot area right now. How do we get going on here. What do I have to pull together? How do I make this happen in 3 to 4 months?

Toby:
Sure. Like I mentioned in segment one, there are really two groups of people out there. There are people who have supply chain design and modeling teams, they have technology in place. For them the focus needs to start looking at not just the long term but the short term. It’s what are some tactical things that I might do? Maybe I pick my top five suppliers and I start to look at those 1500 products that those suppliers are providing to me. I take a look at what are the buying behaviors for those products? What are the stocking levels that I need? You don’t have to do all 100% of your products. You can do this 80/20 rule and find those ones that are giving me the most volume, giving me the most revenue and let’s start at setting the inventory for those. A lot of times you focus off of just these long term major big idea things and let’s focus on some tactical decisions that we can change over a month or two.

Similarly, I guess, if you don’t have a team, you can do that as an outsource project just to get started. To help you justify you know why do I need a technology like that? Why do I need people focused on that? Those are some areas where consulting organizations, technology providers like ourselves that have services arms, we can help you get started by let’s take a swath of a specific group of suppliers, a specific group of products, let’s take a look at that. Over the course of 30, 60 days you can find some cost savings that you can implement right away.

Dan:
Yeah and the savings here can be pretty powerful. You can really reduce your inventories pretty substantially like you said right sizing. Sometimes you’ve got to bump them up to meet demand in other areas but these are pretty substantial savings in a relatively short period of time.

Toby:
Yeah. Multiple millions of cost savings and just about every time that you do one of those projects it justifies why do I need people looking at this all the time. That’s why.

Dan:
If you’re not doing inventory design and modeling now demand segmentation might be the right place to start. Again, I don’t know how you do it without the right tool. Toby, as always, a great discussion, I’m looking forward to part three.

Toby:
Thanks.

Ready to learn more? Click to request a live demo with the LLamasoft team.

Request A Demo