The Inverted W Planning Process
Teams propose, leaders synthesise, teams adjust, leaders resynthesise, orgs distribute with context.
- Difficulty
- Advanced
- Time to result
- ~months to results
- Steps
- 6
- Confidence
- 90%
Stripe's planning process runs neither top-down nor bottom-up but as an inverted W: up from teams, down to leaders for synthesis, back up to teams to react, down again for a final synthesis, then out for distribution. It is preceded by a per-area clarity exercise on exactly which users each team serves, and it is rebuilt from first principles each cycle because a rapidly-growing company outgrows its own planning process.
Origin
David Singleton describing Stripe's annual (deep) and mid-year (lighter) planning. Stripe deliberately designs internal processes from first principles rather than adopting an off-the-shelf system, while learning from other companies — Singleton met Charlie Bell, then running Amazon operations, in his first weeks at Stripe, and much of what Stripe does traces back to those conversations.
Core principles
- 01Design internal processes from first principles for your situation; then go learn from other companies rather than adopting their system wholesale.
- 02A rapidly-growing company will outgrow its planning process, so expect to make sweeping changes rather than incremental tuning.
- 03Planning quality starts with clarity about which users each area serves — the needs of a solo Atlas founder and a Fortune 5 company are wildly different.
- 04Neither pure top-down nor pure bottom-up works: synthesis must happen more than once, with teams reacting in between.
- 05The final downward pass is about distributing context, not orders.
- 06A planning process that people grumble about can still be highly effective; internal NPS is not the metric.
How to run it
- 1
Get explicit about who each area serves
For every product area, state which users you are seeking to serve and hold their needs front and centre. Stripe serves everyone from a founder starting a company with Atlas to Fortune 5 multinationals, and those user needs diverge sharply.
Watch out A team that hasn't named its users will produce a plan that quietly averages across all of them and satisfies none.
- 2
Up — teams surface their immediate thinking
Have teams put forward what they believe the most important things to do are, from their own ground-level view.
- 3
Down — product leaders synthesise a draft company strategy
A group of product leaders gets together and synthesises the most important parts of what came up into a draft overall company strategy.
Watch out This is a synthesis, not a selection: leaders who simply pick winners from the team submissions lose the cross-team strategy that only they can see.
- 4
Up — teams react to the draft
Take the draft back down to teams and ask: if that's where we're making a big push, should that tweak my plans at all? This is the step that most single-pass planning processes omit.
Pro tip This step is what converts a leadership strategy into a plan teams actually own.
- 5
Down — final synthesis
Bring the adjusted plans back up for a second synthesis, producing the plan of record.
- 6
Out — distribute with a lot of context
Push the plan back down so that everyone in each org receives it with the reasoning behind it, not just the conclusions. Context is what allows thousands of daily decentralised decisions to ladder up correctly.
Watch out Distributing the decisions without the reasoning is the most expensive shortcut in planning — every downstream trade-off then gets made blind.
In the wild
Stripe does planning in a relatively deep way once a year and in a lighter way halfway through. Because the company grows rapidly and runs the deep process only annually, the right planning process has changed dramatically and significantly over the years rather than being incrementally tuned.
→ Singleton notes planning does not get the highest internal net promoter score at Stripe, but he considers it genuinely effective — because it is re-derived to fit the company's current scale rather than inherited.
In his first couple of weeks at Stripe, Singleton met Charlie Bell, who then ran operations at Amazon (an existing Stripe user). Many things Stripe now does trace back to that conversation and to conversations with many other companies at similar stage and scale.
→ Singleton is explicit that it was never 'we'll take the thing because Amazon did it' — the practice was to learn broadly, then reason from first principles about what fits Stripe.
Common mistakes
Picking a planning system off the shelf
Adopting another company's system imports their scale, stage and constraints along with their process. Stripe deliberately reasons from first principles about what its users need and how it should therefore plan.
Single-pass planning
Pure bottom-up produces a plan with no company strategy; pure top-down produces a plan teams don't own. The inverted W exists precisely because one pass in either direction is insufficient.
Judging the process by how much people enjoy it
Stripe's planning has a mediocre internal reputation and is nonetheless effective. Optimising planning for popularity leads to processes that avoid the hard prioritisation conversations.
Is it for you?
Best for
Product and engineering leaders at fast-growing companies (roughly 200-5,000 people) whose planning process is being reinvented every cycle and who need a shape for it.
Not ideal for
Small teams where a single planning conversation suffices, or slow-growing organisations where an established annual process can simply be tuned year over year.
From the transcript
“we run a a kind of inverted W process. So, we typically have teams surface what you know, their immediate thoughts are on the most…”
“Then take that back down to teams to figure out well, how does if that's where we're making a big push, should that tweak my…”
“We we think about a lot of the ways that we work internally from first principles.”
“it's really important to focus on who are the users that we are seeking to serve for any given product area”
From the episode
Building a culture of excellence
David Singleton (CTO of Stripe)