Purchase frequency and the annualization error
Durable goods create a recurrent analytical mistake: installed need is treated as annual demand. A household may reasonably own a refrigerator, computer, water filter, camera, vehicle or appliance without buying it every year.
Annualized category spend per owner = Purchase price / Replacement interval
A $2,000 workstation replaced every five years represents $400 of annualized category expenditure per owner before repairs, accessories and financing. A $30,000 vehicle replaced every ten years represents $3,000 per year. The stock of owned goods can be enormous while the annual market is much smaller.
This matters for high-tech projections because technological progress can lengthen useful life as well as increase capability. If a five-year-old device remains sufficient, the replacement cycle expands. Conversely, rapid obsolescence can shorten the cycle, but then the valuation depends on a persistent reason for consumers to replace otherwise functional capital.
The same logic applies to enterprise software when contracts are re-priced downward, to robots whose useful life extends over many years, and to infrastructure whose capital cost is amortized over decades. A valuation model must annualize the transaction, not capitalize the installed base as if it repurchases itself every year.