Excellence Over Speed (The Thought-Experiment Filter)
If you must choose, choose excellence — and kill most experiments before they ever go live.
- Difficulty
- Easy
- Time to result
- ~days to results
- Steps
- 4
- Confidence
- 90%
Rathi's contrarian stance against ship-fast orthodoxy: speed and excellence are two different axes, and when you are genuinely forced to pick one, pick excellence. Most experiments, he argues, should be thought experiments — smart people doing meta-thinking would kill them before they consume company time, because they were obviously going to fail. The framework is not 'never ship fast'; it's a decision rule tied to whether the market gives you a differentiated gap to run into. He uses it as his standard PM interview question.
Origin
Rathi's own contrarian position, stated as 'a preference, not a rule'. He operationalises it as his favourite PM interview question and grades the answer on the candidate's assessment of risk, market and competition.
Core principles
- 01Speed and excellence are separate axes, not a single slider — ideally be good at both.
- 02Forced to choose one: choose excellence. Think more, ship better.
- 03Most experiments should be thought experiments — obviously-doomed tests are a tax on company time.
- 04Ship fast only when there's a clear competitive differentiation or a clear market gap you can occupy.
- 05'Let's ship it and find out' is a tell that the person hasn't assessed the market.
- 06A half-product that isn't differentiated or marketable will not survive a lazy, vain and selfish user.
How to run it
- 1
Separate the two axes before choosing
Stop treating speed vs quality as one dial. Ask which axis is actually constrained here. Only if you are genuinely forced to trade should you invoke the rule.
- 2
Run the experiment in your head first
Before spending engineering time, do the meta-thinking: is this obviously going to fail? If yes, kill it. Rathi's claim is that a large fraction of live experiments would never have launched under this filter.
Pro tip Reserve live experimentation for the tests whose outcome you genuinely cannot predict.
Watch out This is a preference, not a rule — it becomes dangerous if used as an excuse for analysis paralysis or for never testing anything.
- 3
Test the speed case against market reality
Speed is the right call only when there's a clear competitive differentiation or a clear market gap: 'I wanted to launch something even if it is half, no problem — I want to get user feedback and iterate.' Otherwise, excellence.
Pro tip Ask whether the fast version is even viable/differentiated enough to be marketed and taken to users. If not, speed buys you nothing.
Watch out A rushed product that isn't differentiated will simply not work for a lazy, vain and selfish user — you'll learn nothing except that you shipped.
- 4
Use it as a hiring signal
Ask candidates: which products do you decide speed is more important for, and which excellence? Then go into concrete examples where they chose one over the other.
Pro tip Grade on their assessment of risk, and how well they assessed the market and the competitive products in it.
Watch out 'Let's ship something and we'll find out' as a default answer reveals they don't understand viability or differentiation.
In the wild
Rathi's favourite PM interview question is exactly this trade-off. He listens for how the candidate assessed risk, market and competition, then digs into concrete cases where they picked one over the other.
→ It reveals the candidate's frameworks and their core orientation — and whether they are full-stack enough to think about launch, marketing and adoption at all.
Rathi argues that if a company had smart people doing meta-thinking, a large number of A/B experiments would simply never go live, because they were obviously going to fail.
→ Company time is redirected from proving the obvious to building the excellent.
Common mistakes
Defaulting to 'try it and see'
Rathi calls this a waste of company time when the outcome was predictable. It also signals to him, in interviews, that the person hasn't assessed whether the thing they'd ship is even viable or differentiated.
Treating the rule as absolute
Rathi is explicit that it is a preference, not a rule — the point is to be good on BOTH axes; excellence wins only when a genuine either/or exists.
Is it for you?
Best for
PMs and product leaders deciding whether to run an experiment or ship an MVP in a crowded market, and interviewers screening for product judgment
Not ideal for
Genuine land-grab situations with a clear competitive gap, where a half-built but differentiated product should go out immediately for feedback
From the transcript
“think more and ship better most experiments should be thought experiments”
“I think speed and Excellence are two different axis ideally you should be better at both but if you had to choose one choose Excellence”
“which are the products where you decide speed is more important versus which are products where you decide Excellence is what important”
“I look for essentially their assessment of risk”
From the episode
The full-stack PM
Anuj Rathi (Swiggy, Jupiter Money, Flipkart)