In my previous post, I concentrated on the river body of our Lean River System (the balance sheet) to illustrate that there are costs there that play a very significant role in the performance of any enterprise. That inventory and Working Capital are listed on the financial statement as “assets” tends to divert attention from the fact that these entities represent real costs to the business – sunk costs.
When Ohno talked about “cutting costs”, he was referring to both the sunk costs in the balance sheet and the operating costs in the P&L. If I had to guess, I’d say that the former was more important in his mind than the latter. During the early days of TPS, Toyota was trying to catch up to the West but didn’t have the cash resources that were available to them.
I would now like to discuss a few, very common, instances where the traditional cost accounting focus on the P&L pushes for doing the “wrong things right” – local efficiency; but ROI supports doing the “right things”– system effectiveness. System effectiveness, as used here, is a measure of free cash flow per unit cash invested.
In each instance, I will also utilize the Lean River System model to help us to visualize the underlying logic. (See “The Lean River System” series starting here).
For reference, here is ROI (Return on Investment) again:
ROI = Earnings/Total Investment = (Sales – Cost of Sales)/(Working Capital + Permanent Investment)
Let’s look at a couple of typical rationales for overproduction and compare how the P&L and ROI approaches differ from each other.
Suppose demand drops below installed capacity (including labor) for some indeterminate period of time. You want to produce at the new lower demand levels and use the excess labor to pursue improvement activities (the Lean way).
Finance says that it would be better to continue to produce at the higher level to avoid the standard cost labor, volume and overhead variances that would result. Besides (but they usually don’t say it), that overproduction would just be moved into the inventory Balance Sheet account and the monthly standard cost P&L profit margin for that operation would not deteriorate.
But from an ROI perspective, the opposite conclusion would be reached. The ROI analysis would support operating at the lower demand levels.
If you continue to operate at the higher demand level, ROI will decrease due to 1) the higher inventory that would result (larger denominator) and 2) the lower Sales with the same operating costs (reduced Earnings, smaller numerator).
But if you produced at the lower demand level, 1) inventories would remain the same (same denominator) but 2) operating costs would decrease due to the lower input of raw materials. Sales are lower, but Earnings (numerator) would not suffer as much due to the lower operating costs.
ROI will still be depressed at the lower output levels, but better than the overproduction option. Local efficiency suffers but system effectiveness is optimized.
(Note: I observed this problem, after the fact, at one company I had recently joined. I inquired about a rental warehouse that didn’t seem to be generating any activity. I was told that it was rented a few years back during a demand slowdown to hold product that was produced to maintain budgeted volumes. That quarterly income statement probably looked good on paper but the physical product and the warehouse were still there. An “Oxbow Lake” – see River System – Part 4).
If we analyze this situation using the Lean River System model, we will get the same result. Demand has dropped so Sales (the river output) has decreased. But the overproduction approach does not reduce the river input (operating costs) and as a result the volume of the river body (sunk costs) must increase somewhere in the form of inventory (a bigger lake). Lead times will also increase. Both of which are contrary to our basic Lean Principle (see “What is Lean? – Part 1”).
The underlying principle behind TPS/Lean is the systemic creation of the shortest possible lead time for the continuous flow of materials and information in order to generate the highest quality and lowest cost.
We are doing the wrong thing right.
But if we do the right thing (operate at the new, lower demand), we will decrease the river input (reduced material costs) to help offset the corresponding reduction in river output (sales). Sunk costs (inventory) and lead times are unaffected. We still suffer from lower sales but the system remains optimized.
And the second overproduction rationale….
Let’s revisit Taiichi Ohno’s issues at Toyota when he was trying to reduce equipment changeover times and simultaneously reducing production batch sizes (see River System – Part 1 for his words).
Managers and workers were convinced that it was more efficient to run large batch sizes after an equipment changeover. But Ohno knew that while the equipment efficiency would be better with the large batches, the overall system efficiency would suffer.
Fortunately ROI is a system metric. “Optimizing” changeovers by running large batch sizes hurts ROI versus the SMED, short changeover time and small batch size approach. Earnings could be penalized (higher materials cost with no increase in sales) and inventories will definitely be inflated. The numerator gets smaller and the denominator grows – ROI drops!
Thus while local efficiency suffers, system effectiveness is optimized with SMED. (For those of you using OEE as a key performance metric, be careful).
The Lean River System model, while a little more complex, leads us to the same conclusion but with greater visualization detail. Since a changeover is required we must assume that we are producing at least two different product types in the same river system and they cannot be produced at the same time. We must also assume that separate inventories (lakes) exist for each product type somewhere downstream. If not, the customer (river output) would be starved of any products not being produced at that time.
For simplicity let’s assume only two product types (two lakes) are involved. Both lakes are being drained by the customer at all times.
The volume of each lake must, at a minimum, be equal to the volume of downstream flow that occurs during the changeover time period (customer demand). Otherwise the lake would dry up and the customer would be starved during changeover.
In addition, the volume of each lake must also be sufficient to handle the volume of downstream flow (customer demand) that occurs during the time period that the other product type is being changed over and produced.
Thus the minimum batch size of each product type must be at least large enough to restock the given lake, including the ongoing customer demand that occurs during this batch production cycle time. Anything larger than this minimum batch size would unnecessarily increase the lake/river volume (add sunk costs) and lengthen the river lead time. Again, both go against the basic Lean Principle.
Of course, this minimum batch size is theoretical and the actual batch size must contain sufficient safety stock to handle normal variation.
Thus as we reduce changeover times, we should reduce the size of each lake (inventory) to reflect this gain in flow that was previously lost during changeovers. This means we must reduce the minimum batch size in order to capture this gain and minimize sunk costs and lead time. Keep the river optimized!
I think it is worth noting here that while reducing changeover times should lead to reduced batch sizes, reducing batch sizes must, by necessity, lead to more frequent changeovers. Thus reduced changeover time must lead to more frequent changeovers if the river system (and ROI) are to be optimized.
We move closer and closer to continuous flow, which is in line with the basic Lean Principle. This is, of course, contrary to the P&L argument.
Now let’s visit a typical argument against inventory reduction….
Your Lean transformation is progressing nicely. Intra-departmental lead times and batch sizes are down significantly, but the large, inter-departmental, in-process inventories resulting from the past overproduction days are interfering with your continuous flow efforts. You want to lower those in-process inventories significantly.
Finance says we can’t afford to do that because that would require idle labor, and the large labor variances coupled with the overhead variances due to the lower volume, would hurt the standard cost P&L.
But ROI would improve significantly. Inventories would show a large decrease (large decrease in denominator) and Earnings would improve (larger numerator) since sales remained the same (you are selling inventory) but operating costs would be significantly reduced due to the lower material input requirements.
Lean wins! System effectiveness beats local efficiency!
(Note: I experienced this problem while working for a very large Company with very high margins. The Balance Sheet was not important to them. But I experienced the opposite when working for a much smaller firm. They used the cash generated from the inventory reduction to pay off debt. They paid a lot of attention to the Balance Sheet!)
This is a no-brainer using the Lean River System model. You drain the lakes by reducing the river input (material cost) while the output (sales) remains constant. Sunk costs and lead times drop dramatically.
So think ROI if you want to make better Lean decisions. And use the Lean River System model to help visualize the total system. Now you need to get the Finance Department to do the same. (Or you can try to talk CNBC into including “Working Capital Turnover” at the bottom of their “Earning Season” screenshots – I haven’t tried that yet).