9% seems like a large enough number to destroy moral, but too small to make any material differences to the cost base.
Often, "cut fast and deep" makes sense because you're in a difficult situation where both expenses and time are against you. People use all sorts of analogies (e.g. "cutting the fat") but the reality is that if you take out 20-30% of your workforce you will cut back needed functions and you will lose people who are both good and whom you need: there simply won't be enough time, insight or structural ability not to make some mistakes. There tends to be a big impact on morale for a long time after the event - improving morale becomes a major management objective. The upside of this approach is you do it once, it's a big shock internally but then everyone can focus back around the mission.
The alternative approach is a more gradual deflation of your numbers, closing down teams and not-backfilling roles. If you have time then the upsides of this approach are that you can be much more surgical about the changes. Presented in the right way it's going to feel more like a "pivot" with some natural wastage of people that won't work in the new mission. The downside - is that it can feel like a "long march" as successive sets of cuts are made.
Pick your poison as it's both the framework and the implementation!
Reasons why they might have chosen a number of this magnitude?
10-15% annual attrition for a large corporation is fairly normal i.e. losing that sort of experience spread around an organisation isn't particularly challenging (although you would normally backfill some/all). You wouldn't have to do some massive re-org just for this reason (there might be other reasons to do one of course).
In many jurisdictions there are different rules for layoffs over a certain size. So they may be choosing to trigger rules in some areas and not others. Occasionally I have seen companies choosing to "overcut" in a particular jurisdiction in order to take the legal pain once rather than drip feed pain over a long period.
It may also be that they can cut or sell specific areas so that pervasive impact is limited. This might produce a number that's larger or smaller than needed for cost purposes but helps to control the morale issues as many people see no direct impact.
Edit: another very popular reaons is that that's the number that the analysts are expecting. I didn't say it's a good reason, just a common one.
There are plenty of other reasons for ending up with a number up to and including "ip dip sky blue" that I have seen over the years. Generally though, it's tricky and often unpleasant.
Though, Twitter would need to make the layoffs random, force co-workers to escort the laid off off premises, and take away perks for a few weeks... If that's what they were angling for.