A one-point drop in yield sounds small. Going from 96% to 95% is the kind of change that disappears into the normal scatter of a batch record, and nobody opens a deviation for it. But a plant that runs hundreds of batches a year turns that single point into a very large number of tablets and a bill that is easy to see once it is added up. This guide shows how to put a price on it, so that a small trend can be discussed in terms that managers and finance teams understand.
The numbers come from the same supposed batch used in How to Calculate Cost per Tablet. Replace every figure with your own.
The calculation
Yearly cost = Tablets lost per year × Cost per tablet
Equivalent batches = Tablets lost ÷ Tablets per batch
Percentage Yield Calculator, which compares actual with theoretical yield and shows the loss.
Practical Example
In the pharmaceutical industry a product runs 200 batches a year, each with a batch size of 400,000 tablets, so the yearly theoretical output is 80,000,000 tablets. The yield falls by 1 point.
- Tablets lost = 80,000,000 × 1% = 800,000 tablets.
- Equivalent batches = 800,000 ÷ 400,000 = 2.0 batches a year that are paid for and never sold in reality.
- Cost of materials and packaging only (0.04113 CU per tablet) = 800,000 × 0.04113 = about 32,900 CU.
- Cost on a full batch-cost basis (0.04503 CU per tablet, including labour and overhead) = about 36,000 CU.
The API alone is a large part of it so 1% of the 101.01 kg of API used in each batch is about 1.01 kg, so over 200 batches about 202 kg of API, worth around 24,200 CU at 120 CU/kg.
Three ways to value a lost tablet
- Materials and packaging only. It counts only the material that cannot be recovered, and leaves out labour and overhead.
- Full batch cost. Counts everything spent on the batch, spread over fewer good tablets. This is the figure finance usually wants.
- Lost sales margin. If the plant cannot make enough to meet demand, each lost tablet is also a sale that could not be made. At a supposed selling price of 0.09 CU per tablet, 800,000 tablets would be 72,000 CU of sales.
Bigger drops, bigger bills
| Yield drop | Good tablets lost per year | Equivalent batches | Materials and packaging only (CU) | Full batch cost (CU) |
|---|---|---|---|---|
| 0.5 points | 400,000 | 1.0 | 16,451 | 18,011 |
| 1 point | 800,000 | 2.0 | 32,902 | 36,022 |
| 2 points | 1,600,000 | 4.0 | 65,805 | 72,045 |
| 3 points | 2,400,000 | 6.0 | 98,707 | 108,067 |
The loss grows in direct proportion to the drop. The practical message is that a production unit head should look at the yield trend before a limit is breached, and not only react to batches that fail.
Where in the process the loss happens matters
A tablet lost early carries only the cost spent so far. A tablet lost at the end carries almost everything. In the example, API and excipients cost about 0.0308 CU per tablet, while the full cost at the end is 0.0450 CU. So the same one-point drop is more expensive at coating or packing than at granulation. This is why stage reconciliations, such as coating and packing, are worth following separately. The causes at each stage are listed in Yield Loss Points Across Stages.
Try it yourself
Quick calculator (reference only)
Use your own batch count, batch size and cost per tablet. The result is the yearly cost of the yield drop.
Turning the number into action
- Calculate the yearly cost of one yield point for your top five products. It shows where attention pays back most.
- Track the yield by stage, so you can see which stage is moving.
- Set alert limits from your own data, as described in How to Set Yield Limits from Your Own Batch Data, so a trend triggers a look before a batch fails.
- Compare the cost of an improvement project, such as better dust extraction or a tooling change, with the yearly value of the points it can recover.
- Re-check after the change, with the same method, to confirm the saving.
Example: Does an improvement project pay back?
Suppose a dust extraction upgrade on the press costs 15,000 CU and is expected to recover 0.4 points of yield. On the full-cost basis, one point is worth about 36,022 CU a year, so 0.4 points is worth about 14,400 CU a year, and the project pays back in roughly 1.0 years. On the more careful materials-and-packaging basis the yearly value is about 13,200 CU, and the payback is about 1.1 years. Showing both figures makes the case honest and hard to challenge.
Common mistakes
- Valuing lost tablets at the selling price when the plant has spare capacity and the tablets would not have been sold anyway.
- Using the cost of an average tablet for a loss that happens early in the process.
- Counting a yield point as a percentage of the yield, not as a percentage of the theoretical output. Going from 95% to 94% is 1 point, not 1% of 95.
- Adding the materials, full-cost and lost-sales figures together.
- Ignoring the fact that a lost batch also takes machine time and cleaning time.
- Quoting a yearly figure without saying how many batches it assumes.
Frequently asked questions
How do you calculate the cost of lost yield?
Multiply the yearly theoretical output by the yield drop to get the tablets lost, then multiply by the cost per tablet.
Is a drop from 96% to 95% a 1% drop?
It is a drop of 1 percentage point. Measured against the theoretical output it is 1% of output, but measured against the yield it is about 1.04%.
Which cost per tablet should I use?
Use the materials-and-packaging cost for a careful estimate and the full batch cost when finance asks for the total impact. Say which one you used.
Does a small yield change really matter?
For a product with many batches a year, yes. A change that is invisible in one batch record adds up to several batches of lost output over a year.