The PodKnowledge|← All Episodes
Based on Lenny's Podcast data
Lenny's Knowledge Sketch · The Art of Product Management

5 Big Ideas from
Shreyas Doshi

Shreyas Doshi
Former PM: Stripe, Twitter, Google, Yahoo
Creator of the LNO Framework
PM CRAFT
Idea #1, Pre-Mortems

Imagine It Already Failed

"It failed miserably" now work backwards: why? TODAY +6 MONTHS
🐯 Tiger
Will actually kill us
🗒 Paper Tiger
Seems scary, isn't
🐘 Elephant
Nobody is saying it
"Imagine this project has miserably failed 6 months from now. Work backwards, what went wrong?"
  • Post-mortems extract insight after failure. Pre-mortems extract it before.
  • The magic: psychological safety to voice concerns that feel "too negative"
  • Include every function. Alternate talking with 5-min silent writing in shared doc.
  • At end: each person votes for the tiger someone else mentioned, not their own
  • Result: shared vocabulary. Future meetings use "I have a tiger" with no hesitation
Idea #2, LNO Framework

Not All Tasks Are Created Equal, Stop Treating Them Like They Are

L Leverage
Put in 1X, get 10-100X back. Write the PRD for the key feature. Do the strategic review.
N Neutral
Put in 1X, get ~1X back. Routine stakeholder updates. Standard team meetings.
O Overhead
Put in 1X, get far less back. Expense reports. Admin. Process tasks nobody reads.
"I was a perfectionist doing L tasks and O tasks identically. That's why my to-do list was endless and I was stressed, I was optimizing overhead like it was strategy."
  • Reserve your highest energy for L tasks. Let your inner perfectionist shine only there.
  • The same activity (e.g. writing a bug report) can be L, N, or O depending on what's at stake
  • The extra time for L tasks must come from N and O tasks, not from working longer hours
  • Most ambitious PMs fail this, they treat everything like it deserves their best work
Ideas #3 & #4

Three Levels of Work, and Why Execution Problems Are Really Strategy Problems

3 Levels of Product Work
Health Work, Keep the product alive and stable. Fix bugs, pay technical debt, maintain reliability.
Growth Work, Move the core metrics. Activation, retention, monetization improvements.
Delight Work, Create new value. New features, new experiences, new reasons to love the product.

When teams don't agree on which level they're working at, tension looks like an execution problem, but it's a strategy problem.

Execution = Strategy in Disguise
WATERLINEEXECUTIONPROBLEMSTRATEGY PROBLEM
"Most execution problems are actually strategy problems. The team isn't slow or dysfunctional, they're executing a fuzzy strategy at full speed."
  • Symptom: "Why aren't we shipping faster?", Real cause: two teams with different strategic assumptions
  • Symptom: "We keep re-prioritizing", Real cause: no clear conviction on what matters most
  • Symptom: "Lots of conflict between teams", Real cause: teams working at different levels without realizing it
  • Fix: name the level, name the goal, get explicit alignment, before blaming execution
Idea #5, Prioritization

The Common Pitfall: Confusing Effort for Impact

  • PMs naturally talk about what they spent the most time building, not what matters most to customers
  • John Collison (Stripe): "Talk about the product in terms customers understand, not in terms of the effort you put in"
  • Feature effort ≠ customer value. The things easiest to build might deliver the most value.
  • Prioritize by impact on the user's life, not by complexity of implementation
  • Ask: what would make users love this product more? (Jack Dorsey's one question)
Jack's one question In a Stripe acquisition meeting with all the metrics and strategy laid out, Jack Dorsey simply asked: "How does this make our users love Twitter more?" Nobody had an answer.
Contrarian

PM Wisdom Worth Unlearning

A post-mortem prevents the next failure INSTEAD → A pre-mortem prevents the failure. Post-mortems extract insight from wreckage. Pre-mortems avoid the wreckage in the first place.
Work harder when overwhelmed INSTEAD → Classify your work first. Overwhelm is almost always caused by doing O tasks with L task energy. Stop, sort, refocus on leverage.
We have an execution problem INSTEAD → Almost certainly a strategy problem. Get explicit on which level of work you're doing and what the goal actually is before optimizing execution.
Highlight the hardest thing you built INSTEAD → Highlight what most changes the customer's life. Effort and impact are independent. The best feature might have taken 2 days to ship.
Based on Shreyas Doshi's episode on Lenny's Podcast. All ideas on this page are from the episode.Watch on YouTubeFollow @shreyas on X
Every sketch here began as a transcript.

Send us yours and we’ll draw it. $1 a transcript.

Sketch mine →
𝕏︎ X / Twitterin LinkedIn📸 Instagram🔗 Copy link
0:00