Discussion: What key metrics to attack, to optimize your inventory?

Traditionally, it has been understood that to improve customer service you have to have higher level of inventory. This ensures that orders are filled quickly and accurately. But it also means that the value of your inventory is high, to the detriment of your organization as it ties up cash and warehouse space that could be put to other and better uses. So we need to optimize inventory in a way that does not impact your service levels and yet frees up cash. There are multiple ways to achieve this and each one of us uses one or other ways to optimize our inventory and yet keep the service levels at high levels. I used social media to gather inputs from industry experts on this subject so that we can uncover the metrics that they are
targeting and how efficiently they are targeting so that others can take benefit of the best practices.


The metrics that I presented to them for starting the discussion were:
  • Forecast Accuracy
  • Supply Chain Visibility
  • Supplier Lead Times
  • Inventory Turns
  • Inter warehouse transfers
  • Field returns process
  • Accuracy in bills of material
  • Lot sizes
  • Inventory accuracy (system vs. physical)
This triggered the discussion and while some of the experts highlighted the importance of the metrics taken from above list, others expanded the list from their experience. Let me know share the comments from the experts.

I will start with the comments of Michael Hortiatis, who is currently the President of FrozenPhoenix LLC. He emphasized that in addition of the metrics mentioned already, he would like to look at:
  • Buying/payment terms for all purchases: the longer the terms the better
  • Selling terms: the shorter the terms the better
  • Quality of RM and WIP inventory: is your inventory balanced
  • Segmentation of product groups: ensure you are measuring and forecasting correctly i.e. forecast the consumable not the package size especially if the product is sold in different size containers
  • ABC: customize the forecasting technique; an A item (high value small quantity) is different to a C item (low value large quantity)
  • Negotiate with the vendor to carry or postpone manufacturing until the item is required
  • Eliminate complexity in product and/or manufacturing in the design phase: use common components and similar or the same manufacturing equipment or process; both of these strategies will reduce inventory items and changeover times (reduces cost which reduces inventory value)
  • Also, inventory is directly proportional to forecast error and service level which means high inventory for high forecast error & high inventory for high service level
His last comment raises a question, whether all customers need to be treated alike in terms of service levels? Do you still want to keep your inventory level high to keep your C level customers happy or it is high time you revisit your customer strategy now? Food for thought, isn’t it? Optimizing your companies inventory in addition to the many other facet of your business can prove to be overwhelming. Consider using a third-party logistics company to help increase business performance

Enikő Pongrácz, who is an experienced supply chain professional and owns ENPO Ltd., focused on the how part in the discussion and highlighted few very important points as below:
  • Departmental Conflicts: Always, there is a permanent conflict between sales/marketing, finance and production. As a consequence inventory will be a disaster ( bottlenecks in implementing action plans) in all companies, where production planning and eventually procurement/replenishment is under manufacturing, demand and forecast is under sales, eventually supply chain is under finance ( last one very common in smaller companies). All supply chain related functions must be independent from other departments, but cooperating with them
  • $ vs. inventory units ( tons, pcs, cases, etc): As a SC head, never accept a $ measure responsibility for inventory, since this is marketing/product mix depending, which is always changing, and you can not react this change. Yes, the monthly controlling of inventory has to be done in value as well, not only units. According to my experience, the result of RCA was that 80% of the extra inventory is caused by backorders ( production schedule adherence failure) in case of a manufacturing site - and shifts in sales - sales vs forecast differs in product mix in both cases ( manufacturing and trade entities too). This gaps MUST be quantified/evaluated, withdrawn from inventory, to show what would have been the inventory without all quantified failures
  • Inventory turns and DOH - sales vs. MOQ: Measures should be done on actual and plan/forecast quantities as well, ongoing. Just doing a portfolio optimization, you can reduce inventory easily by 10-15%. Again, shared responsibilities - to show marketing and sales, that the profit on some/several products are eaten up by inventory investment - WC or worse, obsolete. Stop selling those slow moving products, where batch/lot can not be reduced or sales can not be increased lets say within 3 month. Add the problematic items on a "portfolio optimization" slide during the S&OP, to keep continuous focus - "items with actual and forecasted sales implying X times vs. average inventory DOH"
  • Sourcing - Nowadays a lot of consignment deals can be closed as well, where only materials consumed are paid. On the other side, doing several post and pre analysis of low cost country sourcing, turned out, that buying locally more expensive, would still be cheaper than holding inventories of X month + emergency airfreight shipment costs
  • Pack Mat and Pallets - These can be hidden items in the pocket, which are influencing inventory significantly (industry depending - the higher according to my experience was in the paper industry, where is high volume / low value ratio). Design separately from trivial items your pack mat and pallets SC flow. What is the standard cost for one way and regular pallets (EURO/ USD)? - normally finance puts them on an average buy price, however one way should be zero std cost and regular should be between 0,5-0,9 * purchase price (used or new = already ones used). Are pallets recovered from customers? If not, are they included in product price or are just a profit decreasing item? etc. Again, having the numbers, after doing the SC part of improvement ( suppliers, 3PL, WHS, etc), co-operation with sales/customers should follow
  • Last but not least - different SL's for different products and different customers is just the start up tactic of CS level improvements at the beginning. Reaching above 95% CS level, there should not be any differentiation; otherwise the focus on the cause is lost. Part data does not give the same RCA% as whole database

Rohan Patkar, who is a Senior Manager at Algorhythm Tech. took us to wards a new direction. As per him, Inventory optimization can carried out at two levels - strategic & operational. Strategic level includes changing the Bills of distribution, changing transportation modes, changing the review periods, reducing cycle times etc.

Operational level includes:
  • Correct modeling of demand & supply variability (many inventory modeling approaches start by assuming a normal distribution of the demand and supply variability which for him is not correct)
  • Having differential service level policy for different products & customers
  • Mapping the interdependencies between echelons in the supply chain. Considering the interaction between downstream service levels and the upstream service levels
Global Supply Chain Director at Vertellus Specialties, Luiz Gustavo Araujo, mentioned, “In my view, cycle time is an appropriate metric for inventory optimization. If you want to reduce inventory, drive cycle times down. Cycle time can be broken down for each step of the process and if you are able to map out the different cycle times, you can develop a plan o attach to reduce waste in your process. In many cases, cycle times to replenish a warehouse are a big one to start. Replenishment lead times are often overstated for the sake of being conservative and we still go ahead and add safety stocks. By identifying the different cycle times in the process, you can eliminate waste.”

Matt Hunter, who is a senior purchasing manager at Drummond Company told us that since they are in a capital equipment intensive business and are very sensitive to down time metrics, by considering carrying cost, purchasing cost, usage, lead times, and most importantly, criticality of the inventory item, they can optimize service levels or fill rate. In other words, the more critical the item, the higher the service level. Critical items include items that will completely stop production where no work-around options are available. Bottleneck operations are also an appropriate focus in flow-based operations. It seems a good way to optimize your inventory at least for the non critical items.

Bob DeRosier, who just moved from Johnson and Johnson as senior eCommerce analyst, feels that cycle time reduction, better forecast accuracy and alignment are the key metrics one should target to optimize inventory. He says that it is the forecast accuracy that determines how much of a certain item to make and the higher the accuracy error, the more inventory you will have, which is straight forward. From an alignment perspective all owners of inventory production should be aligned to the same metrics. Sometimes this isn't the case and more (or less) inventory could simply be because different groups are trying to attain different numbers.

He recognizes cycle time as another focus area in this regard and added that as the cycle time of inventory is reduced, turns will increase and inventory will be optimized.

In addition, he mentioned SKU rationalization as another area here. He asked us, do a lot of companies keep those C SKU's just because they move slowly? Maybe this is a little off the intent of the topic but in order to optimize inventory and drive costs down this is something that definitely needs to be looked at as those SKU's take up space and offer little to no value. Fair point this, isn’t it?

Finally, I will close the topic after mentioning the remarks of Hugh Kinney, who is currently the Principal at Altus Ventures and said that, “I would add demand and lead-time variability as they are important inputs for optimization algorithms. Computation and measurement of inventory carrying cost is also important. Additionally a very important point from him - The main element that is not addressed is the reward system for the players involved in the process. If sales and manufacturing have no incentive to drive inventory reduction or optimization, then it will not happen. If you can tie inventory carrying cost to the commission plan for sales and the bonus plan for manufacturing management, you have a chance to optimize, otherwise, forget it. Try subtracting carrying cost from gross margin when computing sales commissions and adding carrying cost to manufactured unit cost when computing the bonuses of the manufacturing team. Metrics are good but they will only drive behavior when associated with some type of reward or incentive. The rewards do not have to be monetary but some reward or incentive such as recognition, promotion, celebration, etc should be incorporated for best results."


Well, everyone contributed with their cents in the discussion and due to space constraint and similarity of points, while I did not post comments from all but would like to thank them for sharing their wisdom and experience.


  1. I get alot of questions about inventory turns because they can be hard to understand. Even harder to understand or model is how inventory turns affect cash flow. However, the entreprenuer who measures turn and appreciates the connection is much more likely to prosper in the long run... Read more at http://www.mrmissionposible.com/blog

  2. I think there are a lots of metrics to control inventory. We shouldn't pay too much attention to particular metric but trying balancing various metrics instead.

  3. Right forecasting tools are a must these days as inventory carrying costs are increasing and are pinching more and more to business. A nice article though...


Your thoughts are welcome