Matt Mullenweg recommended this post and I love it! I summarize as a means to learn, and here’s my learnings. —JM
Where To Invest
| Return on Investment Comparing Stocks S1(t,m,p) vs S1 (t,m,p) → more money Comparing Projects P1(t,m,p) vs P2 (t,m,p) → more customer impact Maximize customer impact/unit resource. | Time Value of Shipping Pa: Build Resulting in big customer value → launch slower Pb: Build Resulting in smaller customer value → launch faster Measured over a longer time, Pb is 👍. |
| Time Horizon If time span for impact is large THEN → comparatively speaking, Pa is 👍 If time span for impact is short THEN → comparatively speaking, Pb is 👍 The allowable horizon of time governs all. | Expected Value Eva = Pav,success * Psuccess + Pav,failure * Pfailure Evb = Pbv,success * Psuccess + Pbv,failure * Pfailure Pick the higher Ev is b- |
Designing and Scoping
| Working Backwards Conventional Start with → devise solution outcome Inversion Start with outcome → devise starting point Working forwards doesn’t optimize impact. | Confidence ↝ Speed and Quality Importance of Quality ∝ High confidence in problem + solution → validated thru customers Importance of Speed ∝ Low confidence in problem → no customer validation For quality, no shortcuts; for speed, test fast. |
| Solve the Whole Customer Experience Don’t overfocus on the in-product experience. → Design → Design the support experience. Support experiences engender trust. | Experiment, Feature, Platform Experiment / hi-speed, Learn how to solve Feature/ speed = quality Solve customer problem Platform / lo-speed, hi-quality Enable more features |
| Feedback Loops Equilibrium is easily upset for good and bad. | Recursive Feedback Loops Flywheel effect in a system is 👍. |
Building and Iterating
| Diminishing Returns “When you focus on improving the same product area, the amount of customer value created over time will diminish for every unit of effort.” | Local Maxima “Incremental improvements creates no customer value at all, forcing you to make a step change in product capabilities.” |
| Version Two is a Lie “Make sure that whatever you ship is a ‘complete product’ which, if it was never improved again, would still be useful to customers for the foreseeable future. | Freeroll “When the current UX is so bad that by making any reasonable change based on intuition is likely to make it much better.” 🎲🎲 |
| Most Value Is Created After V1 “You will learn the most about the customer after you launch the product, don’t waste the opportunity to build on those learnings.” | Key Failure Indicator (KFI) “Pairing your KPIs with metrics you don’t want to see go in a certain direction, to ensure you’re focused on healthy growth” |
