Build the Right Thing
=====================

Paul Murston

- have idea
- build it
- put it online, market it
- wait for sales, users ...

- give up (?)

All guesswork based on assumptions

Pain

- invest time, money ideas
- pour heart and passion in
- then fail!

Dream

- happy users/customers, love it

"Not worth it unless you have users/customers"

Happy User Factory

- find where they hang out
- what they say, what they do, what they make, what they **use**
- what job are they trying to do? what pain do they have? what gains do they hope for?

Kill Pain or Accelerate Gain
----------------------------

Turn our users into badasses!

problem hypothesis
solution hypothesis - how does project make it better

can I fake it before I build it?
can I demonstrate the gain before building it?
if I have to build it, what is the smallest thing I can do to get traction?

Scenarios. Use cases.
What is the central thing that delivers value
rapidly prototype

NOT about UX, scaling, security, pythonic cool
IS about **being useful**

Tech stack
Product katas
Practice many times
Prune your stack - optimise for expressiveness and productivity
Especially productivity - cookiecutters, scaffolding, "production capability"
Don't do everything - deliver key value, then build out

Get people whose problem you attempting to solve to have **skin in the game**
eg classically money - but also reputation, time
once they're committed, they'll be honest with you (before they'll just be nice)

use available resources - mgt science "effectivity" - what resources, USP do you have
100% new market - almost bound to fail
changing 20-40% - you have solid blocks to build on

**steal ideas** - from other domains

"A series of experiments are worth more than a good idea"

"A good plan violently executed now is better than a perfect plan next week" General Patton - there are always excuses for not going now

Many decisions are reversible - make those quickly

tempo is your killer feature (if you are small startup/one person)

get deployed quickly
**no excuses**
start the conversation