Here’s a profile of my approximate yearly expenses, as generated by the budget-projection library I’m working on:
Total expenses: -26763.315821
Profile of expenses:
rent+utils -14160.00 -------------------------------
lunch -2920.00 ------
dinner -2920.00 ------
breakfast -1460.00 ---
unforeseen -1414.96 ---
wing chun -1200.00 --
going out -780.00 -
laundry -624.00 -
mta -432.00
transportation -360.00
rackspace -132.00
spotify -120.00
haircut -120.00
netflix -96.00
I would imagine that your profile looks similar. Clearly, something like Netflix (or any other service recurring monthly, save rent) isn’t the bottleneck here. As in software optimization, you have to look at the inner loops — the costs you incur most often — and reduce those. For example, the food that you eat, or the coffee you buy. A cappuccino per day at $3.50 (cost of a decent one in Brooklyn) will cost you nearly $1300 per year.
This projection is after taking into account that I’ll be cooking most of my meals at home. When I was eating out more frequently, food was (predictably) more expensive. I’ve estimated lunch/dinner at $8 a pop, which may sound slightly optimistic, but is attainable. Today’s lunch consisted of fried arctic char and a batch of asparagus ($8 on the nose). Tonight’s dinner consisted of two eggs, three sausages, peppers, and mushrooms (a whopping $3.80). I have cheated slightly: I ate strawberries and raspberries before dinner.