Percentages may be "de-humanizing" but they are much better for understanding the relevance and severity of layoffs. If your local bakery lets 20 people go it's much more likely to be a sign of significant issues with the business than if Apple lays off twenty people (which wouldn't even be news).
In fact, when reading news stories you should _always_ look for the denominator (and be on guard when it isn't provided), lest you be mislead by emotionally framed numbers floating in the void.
I kind of wish they did both... percentages are more helpful in understanding the impact (in this case to the company). But you're right in the case of layoffs a count is more meaningful... although i'd argue whether it's 1 or more the impact to the individual is the same and usually very negative.
The article says this is an estimate based on an "about 10%" figure:
> Implied was that they would be doing this without the help of an undisclosed number of newly former employees. That number has been guesstimated at around ten percent of the workforce – about 12,500 people – as part of an effort to get the overall workforce from 120,000 to below 100,000.
For press release or to investors, the percent is the number that matters.
To the internal communication to the employee losing their jobs or those that have to keep the torch going, the number is usually better.
With that said, seems like it's part of a larger downsizing plan: "That number has been guesstimated at around ten percent of the workforce – about 12,500 people – as part of an effort to get the overall workforce from 120,000 to below 100,000."
I agree. 10 is 10% of 100, 1,000,000 is 10% of 10,000,000. They're mathematically equivalent. Why do I care if it's 10 people, or 10,000,000 people, or if it's just quoted as 10%? These are all abstract numbers, numbers are not matters of humanity.
In fact, when reading news stories you should _always_ look for the denominator (and be on guard when it isn't provided), lest you be mislead by emotionally framed numbers floating in the void.