Simple answer: net marginal customer revenue.
I'm not sure if that is an accounting term but I've seen it used in similar situations to describe how the business model works, or doesn't work.
It goes kind of like this:
Lets say that 1 SRE engineer, 1/4th developer, 1/3rd customer support agent, and 1/10th of a manager can support 10,000 users. (all made up numbers of course for this example). You can't really cut up manager's in 10ths so you really need 10x this to have 10 SRE engineers,4 developers, 3 customer support agents, and 1 manager.
So that is 100,000 users. But the cost of all those people don't bring in enough revenue from those users.
But there is a "scaling" effect, such that if you have twice that many people, instead of 200,000 users you can support 225,000 users. Four times and instead of 450,0000 users you can support 500,000 users. So the trick is to find the point in the model where you have enough users to make enough revenue to pay for all those people, and the hardware that supplies the service. At that point you're "break even", and then if you can push past it, you start generate cash above what you need, or net profit.
The more you can grow, the more net cash you can generate, and the "value" of your company, is typically a function on your earnings before you start paying taxes, and costing depreciation (EBITDA).
A web business like Twitter's (or Facebook's for that matter) then can be valued based on how many users they have and the revenue per thousand user metrics they can drive minus the cost per thousand users.
If you stop growing before your generating net income, you die. So there is a minimum you have to hit. And if you're well managed, you are constantly trying to minimize costs and maximize revenue to get a bigger income factor.
Here is another issue, generally, like with "blade computers" you get the most money by having just enough people and resources to handle all of your load. When you anticipate growth, if you build out more resources and that growth doesn't come, your stuck with the costs. Like having a computing blades chassis with just one blade in it. A really expensive computer. So you back off and lay off those folks you hired to get your costs back under control so you don't lose a ton of money.
Again, a well managed company will attempt to grow without hiring resources right up until they can't grow any further, and then hire. So that when the next tranche of people show up, you are ideally already net positive on income, just not by as large a margin as you used to be.
At the end of the day you're managing costs, user growth, and revenue per user. If you can't juggle all three you run a high risk of becoming an acquisition by someone who can.