Your Factory Was Built on Assumptions. What Happens When They Change?

Bosch just took a €270 million impairment because EV demand didn’t arrive on schedule. The machines still work — what changed were the assumptions underneath them.

Key takeaways

• Every production asset rests on assumptions about demand, mix, cycle time, yield, labor and margin. The asset lasts decades. The assumptions rarely do.
• The Assumption Gap is the distance between what the business expected when it approved capacity and what is actually happening on the floor today.
• Utilization alone does not tell you if capacity creates value. Manufacturers need to separate Installed, Available and Economic Capacity.
• Underused capacity is not automatically bad capacity. Before cutting it, ask what else it could make.
• The earlier you see the gap, the more options you keep.

What is the Assumption Gap in manufacturing?

The Assumption Gap is the difference between the operating assumptions used to justify a manufacturing investment and the conditions the factory actually faces now. It covers demand, product mix, cycle time, yield, labor, supplier lead times and margin. When the gap grows unnoticed, capacity that looked profitable on paper slowly stops earning its keep.

Every factory starts as a set of beliefs: how much customers will buy, which products they will want, how fast a line will run, what labor will cost, what yield will be, how long suppliers will take. Those beliefs turn into very real things — machines, lines, people, inventory, warehouse space, supplier contracts, buildings and capital.

Factories last for decades. Assumptions don’t.

What does a €270 million assumption look like?

Bosch just showed us. In its first half 2026 results, the company said profitability was hit by €270 million in impairment charges on its production facilities, because global electric vehicle adoption lagged behind earlier expectations. Its operating profit margin fell to 4.6% for the first six months of 2026, down from 5.1% a year earlier.

It helps to remember where those expectations came from. In 2024, Bosch estimated that 70 percent of new cars in Europe would likely be purely electric by 2030, with 40 to 50 percent in China and North America. That forecast was reasonable at the time. Capital was deployed against it. Then the market moved at a different speed.

Think about what an impairment actually means. The machines still exist. The buildings still exist. The production capability still exists. What changed was the economic assumption underneath them. In practical terms, Bosch wrote down the accounting value of certain factories and equipment because they are likely to be used less than planned.

Bosch is not alone: Ford announced a $19.5 billion write down tied to its EV operations, GM took $7.6 billion in charges in the second half of 2025 to reduce EV capacity, and Stellantis announced €22.2 billion ($26.2 billion) in charges — the largest yet by any global automaker. Reuters puts the industry total at roughly $55 billion.

Notice what sits inside GM’s number. About $4.2 billion of its January charge was expected to be cash, mostly from terminating or restructuring supplier contracts made when EV sales were expected to grow faster. That’s a supplier commitment built on last year’s forecast, at scale.

Why most manufacturers never see their Assumption Gap

Large companies announce impairments. Most manufacturers experience the same problem quietly:

• An expensive machine runs at 43% utilization.
• A warehouse holds inventory tied to demand that never showed up.
• A line is configured for a product mix customers no longer buy.
• A second shift was built around volumes that disappeared.
• A supplier commitment reflects last year’s forecast.
• Labor is scheduled to the wrong production pattern.

Walk that 43-percent-utilized machine’s aisle and it won’t look broken. An operator is running parts, the light stack is green, and the supervisor has no reason to flag it. What doesn’t show up on a walk-through is the second shift built for a volume forecast that never arrived, or the fact that nobody has recalculated what “good” utilization should even look like for a product mix that shifted two years ago.

Nobody issues a press release for any of it. The cost shows up a little at a time, in margin, absorption and working capital.

This isn’t a niche issue. U.S. manufacturing capacity utilization fell to 75.7 percent in August 2026, 2.5 percentage points below its 1972 to 2025 average, according to the Federal Reserve. Roughly a quarter of the country’s factory capacity is sitting idle on any given day. Some of that is healthy slack. Some of it is an Assumption Gap nobody has measured.

The two gaps: the Gap and the Assumption Gap

At TransLution, we talk a lot about the Gap: the distance between what your ERP believes is happening and what is actually happening on the floor.

• Syspro says the material is there. The operator cannot find it.
• The routing says the job takes 18 minutes. The floor takes 24.
• The ERP says the machine is available. Maintenance knows otherwise.
• The system says production consumed 500 units. Physical inventory disagrees.

The Gap matters because management makes decisions from the system while reality happens somewhere else.

The Assumption Gap sits one level higher, and it can be even more expensive:

None of those differences looks catastrophic alone. Together they can change the economics of an entire plant. The two gaps feed each other — if your system cannot tell you the real cycle time or the real yield, you cannot tell whether your investment assumptions still hold.

Every production asset has a business case. Who is still checking it?

Every significant manufacturing asset started life as a spreadsheet. Someone made assumptions about demand, revenue, volume, utilization, labor, yield, material cost, cycle time, margin and return on investment. Those assumptions justified the capital.

Then the machine is installed, and the business case gets filed away. Management watches whether the asset is running. Far fewer organizations keep asking the more useful question: are the assumptions that justified this asset still true? That’s the question Operational Truth should answer.

How to find the Assumption Gap: the Big Five of Operational Truth

TransLution’s Big Five of Operational Truth give manufacturers a simple structure for testing capacity against reality.

1. What do you have?

Real, usable capacity, not nameplate capacity — capacity after downtime, maintenance, labor availability, changeovers, quality losses, material shortages and actual cycle times. Which assets are genuinely productive? Which are underused? Which are overloaded?

2. Where is it?

Capacity only has value if it exists where demand needs it. One plant can be constrained while another has idle machines. One line can be overloaded while its neighbor runs at half its planned rate.

3. Who is using it?

Which customers and products consume the capacity? A machine can look busy while producing low margin work. Another can look idle because demand shifted away from the products it was designed for. This question connects utilization to business economics.

4. Is it where you thought it would be?

This is where the Assumption Gap becomes visible. When the asset was approved, what demand, mix, utilization, yield, cycle time and margin did the business expect? Compare those numbers with today.

5. What is it costing you?

Look well beyond depreciation: idle labor, poor overhead absorption, unused floor space, tied up working capital, and capital that could be earning a return somewhere else.

Now the conversation has moved from “Is the machine running?” to “Is this asset still creating the value we expected?” That’s a much better question.

What is the difference between installed, available and economic capacity?

Installed Capacity is what the equipment could theoretically produce. Available Capacity is what the operation can actually produce after real world constraints like downtime, labor and changeovers. Economic Capacity is what the business can produce at an acceptable return, given current demand, cost, mix and margin.

A machine can have plenty of Installed Capacity and even Available Capacity. If there is no profitable demand for what it makes, its Economic Capacity may be close to zero.

The reverse is also true. An underused asset can hold significant Economic Capacity if it can be pointed at a higher value product. Utilization alone cannot tell you which situation you are in.

Should manufacturers cut underused capacity or repurpose it?

When demand falls, the instinct is to cut: close the line, sell the machine, reduce the shift, write down the asset. Sometimes that’s exactly right.

But the better first question is what else the operation could make. We call this capacity repurposing.

The auto industry is doing a version of this right now. GM responded to slower EV demand in part by pivoting some assembly capacity from EVs back to internal combustion vehicles. Stellantis reintroduced the Jeep Cherokee as a hybrid and brought back the Hemi V8 in its Ram pickups.

The original assumption was wrong. The plants still had value. The value moved.

The same logic applies to a midsize manufacturer. Picture a line running at 52% utilization. The traditional view says you have 48% excess capacity, which is a cost problem.

Now suppose that equipment can produce another product with limited retooling, that product carries a higher margin, and a new customer needs capacity now while building a new facility would take 18 months. The same idle hours become available productive capacity and a growth option.

Nothing physical changed. What changed was management’s understanding of the options.

Why investment assumptions should have expiration dates

Here’s a simple management practice hidden in all of this: when a major investment is approved, record the operational assumptions underneath it, not just the expected ROI — demand, mix, utilization, cycle time, yield, labor, material cost, supplier performance and margin. Then compare those assumptions with reality continuously, not once a year.

Imagine management seeing this on a single screen:

Now the discussion is useful. Something changed. The next question is what to do about it.

That comparison is only possible if the “actual” column comes from the floor as work happens. Standards in the ERP won’t tell you the cycle time has drifted to 21.4 minutes. Scans, job transactions and labor records captured on the floor will.

Predictability = Visibility + Control

This is the principle behind everything TransLution means by Operational Truth.

Visibility means understanding what is actually happening. Control means seeing it early enough to still have choices.

Demand can drift down for eighteen months before management notices, at which point options are already limited. Mix can shift while the plant keeps buying material and scheduling labor around the old assumptions, and the gap gets expensive. Cycle time can creep up while the standard never changes, until capacity planning becomes fiction.

Seen early, the same drift leaves room to repurpose capacity, adjust staffing, retool for a different product, or simply avoid the next capital request.

The payoff from real time floor data is often faster than people expect. At Lancewood, updating the production plan used to take three to five days in Excel; with stock data pulled in automatically through TransLution, it now takes about three hours.

A&A Electrical cut warehouse costs by up to £200,000 a year. Time creates options, and accurate data creates time.

Don’t wait for the impairment

Bosch could put a €270 million figure next to capacity whose economics changed. Most manufacturers will never see it laid out so neatly. It shows up instead as lower margins, more inventory, missed commitments, another capital request, and a nagging sense the factory isn’t performing the way anyone expected.

So the question isn’t whether the machines are running. It’s whether the assumptions our factory was built on are still true. And if they’re not, the real follow-up is what else the factory we already paid for can do.

Every factory is built on assumptions. The best manufacturers notice when those assumptions change, early enough to change with them.

Want to see where your Assumption Gap sits? Start with TransLution’s 5 Question Audit, or

request a custom demo to see real time floor data flowing into Syspro.

Frequently asked questions

Sources and further reading

• Global Banking and Finance Review (reporting Reuters): Bosch Profit Margin Falls on Weak Car Production, One off Charges
• Finimize: Bosch’s EV Reality Check Is Showing Up in Margins
• Bosch Media Service US: Bosch is banking on innovations, partnerships, and acquisitions
• Axios: Stellantis’ $26 billion write down adds to auto industry’s EV losses
• Autoblog: Major Automakers Have Written Off $55 Billion After Overestimating EV Demand
• CarPro: GM To Write Off $6 Billion In Electric Car Scale Down
• Nasdaq: EV Write Offs Rise, Yet One Auto Giant Is Doubling Down
• Federal Reserve Board: G.17 Industrial Production and Capacity Utilization
• TransLution Software: You Can’t Cut What You Can’t See
• TransLution Software: Your Inventory Report Tells You What You Have. It Doesn’t Tell You When Production Stops.
• TransLution Software: You Can Have Too Much Inventory and Still Not Have What You Need
• TransLution Software: How A&A Electrical cut warehouse costs by up to £200,000 a year