financial literacy & unit economics

one curve, three lenses, and the discipline of never letting a number float free of the business it describes.

the curve everything sits on

akshay's rule for the week: no concept floats. every term must sit somewhere on this curve, and most terms change their answer depending on where you stand.

everything is a hypothesis

the machine works. feed it.

the machine runs. tune it.

the market moved. now what?

same questions. answers flip by stage.

vodafone idea and lenskart walk into your office

three decisions for each company: a key hire, an investment call, a financing choice. same three questions, two very different spots on the curve.

the point you should be able to say out loud by 15:15: your answers flip because the stage changed, not the logic.

the anatomy of one chai

what swiggy spent to get you

the first-order discount, the free delivery, the ads that followed you for a week: add it up and your first biryani cost them several hundred rupees. that bill has a name, CAC, and a question attached: what are you worth to them over the years?

hold that question. it has a name too, and it's next.

the barber's first-haircut discount

half price on your first cut loses him money that day. but a customer who stays comes back every month for years. the discount is CAC; the years of cuts are LTV; and the number of visits until he's made the discount back is the payback period.

the whole subscription economy is this barber, at scale.

the 200-plate wedding order

the tent, the cooks, the transport are fixed: paid whether 50 plates serve or 500. each plate's ingredients are variable. somewhere in that 200-plate order is the plate where the fixed costs are finally eaten, and every plate after it is the caterer's actual profit.

quote a wedding without knowing which plate that is, and you're gambling, not catering.

the FD proof

₹100 in a fixed deposit at 7% becomes ₹107 next year. so ₹107 arriving next year is worth exactly ₹100 today. that 7% is a discount rate: the FD, running backwards. every discount rate is just this, with more risk in it.

the second-outlet decision

₹11.2L of today-money coming in, against ₹10L going out: NPV is positive, the outlet clears. teaching estimates, directionally right.

the same decision, as a rate

IRR is the discount rate at which that NPV hits exactly zero. if the project's IRR beats the rate you demand, it clears the bar. NPV answers in rupees, IRR answers in percent: two views of one question.

tomorrow's question, then: where does the bar itself come from?

choosing between real investments

a new outlet, a new machine, a new market: rank every candidate's NPV against the bar, fund what clears it, starve what doesn't. the discipline is identical at every stage of the curve. only the stakes change: a startup bets the company, a mature firm bets a percentage.

why equity is the expensive money

the blend is the bar

take a monday maturity company: four-fifths funded by equity expecting say 14%, one-fifth by debt costing say 7% after tax. blend by weight and the company's money costs roughly 12.5%. that number is the bar from yesterday: any project earning less than the money funding it destroys value while looking busy.

why vodafone idea can't borrow, and asian paints won't need to

lenders lend against predictable future cash. decline has none it can promise, so no one lends to Vi at a price Vi can survive. maturity has plenty, which is the irony of credit: debt is cheapest for exactly the companies that need it least.

one fund, one decade, one pipeline

when a VC pushes you on timelines, this pipeline is why: your exit has to land inside their fund's life, not just inside your dreams.

the fed and the repo decide how much dry powder exists

cheap money fills LP pockets and LP pockets fill funds: that was 2021. rates rise, the chain runs backwards, and the funding winter you traced in week 3 arrives. the pipeline on the last slide is plumbed straight into the weather.

future cash, dragged to today, summed

yesterday's discounting, industrialised: forecast the cash, discount each year at the bar, add it up. the answer is a value, and every input is an argument.

where most of the number hides

that dashed box, everything after year five, routinely carries the majority of a DCF's value. which means most of the valuation rests on the years you can see least. that should worry you every single time, in your own models most of all.

what would someone actually pay?

the shortcut the market runs on: price = earnings times what buyers currently pay per rupee of earnings for similar firms. PE and its cousins skip the forecast entirely and ask the neighbours instead.

four companies, two methods each, one half-day

teams take one company each and build the valuation both ways: a DCF you can defend line by line, and a multiples view honest about its comparables. all figures on the next slide are teaching estimates, directionally right; your job is the reasoning, not the decimal.

ready to meet them?

press to meet them

your mission business, three lenses, mentors in the room

bring your numbers, whatever state they're in. every business gets the same treatment: where does it sit on the curve, does it make money on one, and what does its money cost? mentors circulate; nobody presents, everybody works.