three half-and-full days on the only question that matters once building is free: what do you refuse to build, and can you say why?
Vaman Sharmamentor · product management
cold open
building is cheap now. this module is about the expensive part: deciding.
refused so far01
the first discipline
fall in love with the problem, not the solution
you arrived with an idea. it feels like the point. it is not: it is your first guess at a problem you have not finished describing. founders who fall for the solution defend it. founders who fall for the problem keep changing the solution until the problem gives way.
the test: if the problem were solved tomorrow by someone else's product, would you be relieved, or disappointed? relieved means you love the problem.
refused so far02
jobs to be done
nobody wants a drill
refused so far03
case · freshworks vs zendesk
simplicity is a feature, and it is expensive to hold
zendesk sold configurability: every workflow, every rule, every integration, priced and built for a buyer with an admin team. freshworks, out of chennai, pointed at the business that has no admin team and wants to answer customers by friday. same job, opposite answer to "should we add this?"
the smaller buyerSMBs and mid-market: businesses who count the setup week as a cost, not an investment
setup in daysthe product promise was time-to-value, which is only possible if you keep refusing features
listed on nasdaq, 2021an indian product company, built on saying no to the enterprise checklist. directional
refused so far04
the one-user rule
build for one real person
name themnot "small business owners". one person, with a name, a shop, a phone and a bad afternoon.
you can call themif you cannot reach your user in one call, you are guessing and calling it strategy.
they decide the tiesevery feature argument ends with: does this help them, this week? "everyone" cannot break a tie. one person can.
lab a · this evening
find the job
take the product you think you want to build, or one you use daily. write the job its user is actually hiring it for: the shelf, not the drill. then find the job they are hiring the alternative for, including doing nothing. the gap between those two is your only real opening.
mission 11 · briefed tonight · submits with the ai mission, tue oct 7
the problem brief
one real problem, one real person who has it, and the evidence that it is worth a term of your life. this brief seeds your PMV in term 2: write it about a problem you would spend a term on.
submits alongside the ai mission on oct 7. the same business can serve both.
day 2 · tuesday
everything you don't build is a decision.
refused so far05
the working tool
value against effort
refused so far06
case · tata neu
the super-app maze
one app for flights, groceries, electronics, medicine, payments and loyalty, assembled from businesses that were each already good. the logic was the group's: every unit gets its place. the user's logic subtracts: I came to do one thing, and now I am choosing between nine.
org logic adds. task logic subtracts. when the home screen looks like the org chart, the refusal never happened.
refused so far07
case · upi
the single QR
the easy build was the one every wallet had already made: our app, our code, our merchant, our walled garden. the refusal was to let any app pay any app, on one code the shopkeeper prints once and never thinks about again.
every player gave up the thing they thought was their moat. the market that opened was bigger than the garden any of them were guarding.
refused so far08
the smallest thing
build, measure, learn
refused so far09
case · pieter levels
a company of one
no co-founder, no funding, no team, no roadmap beyond the next thing, products built and shipped in days and killed just as fast. the discipline is not the speed. it is the scope: every product stays small enough that one person can hold all of it, and anything that would need a second person is not built.
most founders refuse features. he refuses the things that would make features necessary.
labs b + c · fused · one artifact out
kill the roadmap, then spec the one thing
first · kill ityour roadmap, cut to one linelist everything you intended to build. put each on the 2×2. then delete every row but one, and write the reason each deletion is safe for your one user.
then · spec itthe one thing, specifiedwhat it does, who it is for, what it deliberately does not do, and the one question shipping it will answer.
one page out, and the deleted rows stay on it. they are wednesday's defence.
build lab · this afternoon
this afternoon you ship it
the spec from the last lab, built. use whatever gets you to a working thing fastest: AI-assisted building, no-code, a form and a spreadsheet behind it. nobody is scored on the stack. you are scored on whether a real person can use it tonight.
tonight
one real person uses it. not a friend being polite: a person with the problem.
day 3 · wednesday
demo & defend.
the format
three minutes building, three minutes refusing
first · 3 mindemo the ONE thingworking, on screen, doing the job for the user you named. no slides about it.
then · 3 mindefend THREE refusalsthree things you did not build, and why each one waits, tied to your one user.
thenpanel questionsthe panel pushes on the refusals, not the demo. the build is the easy half.
running order and per-fellow slot times come from the scheduler sheet on the morning.
the standard
"I didn't have time" is not a refusal. "my user's real problem is X, so Y waits" is.
the ledger, full · nine refusals in three days
what the module actually taught
01
the solution you already lovethe one you arrived with, before you knew the problem
02
the drillthe feature as the customer asked for it, instead of the job underneath
03
configurabilityfreshworks refused the settings maze zendesk sold as power
04
"everyone"the user who does not exist, in place of the one who does
05
the thankless quadranthigh effort, low value. the work that feels productive and is not
06
the org chart on the home screenthe refusal tata neu did not make
07
the walled gardenevery app its own QR. upi refused it, and the market opened
08
everything that does not teach youanything the smallest version does not need to answer its question
09
headcount, funding, scopelevels refused all three, and shipped anyway
seeing clearly is choosing what not to do. that is the whole term, in one column.
the hinge
term 1 taught you to see. term 2, you build.
behind youtwelve weeks of seeingeconomics, selling, decisions, money, organisations, leadership, voids, intelligence. every one of them a way of looking at a business from the outside.
aheadthe PMVyou build one thing, for one person, and defend it for a term. your problem brief is the bridge: it is already written.
close
tomorrow the term ends, and your october letters open.
the school of future