Three Divergent PRFAQs (The Power of Three)
Write three fully-thought-through, divergent press releases — then recommend one.
- Difficulty
- Advanced
- Time to result
- ~weeks to results
- Steps
- 5
- Confidence
- 95%
Rathi's operational upgrade to Amazon's working-backwards process. Instead of working backwards from one customer value proposition, you work backwards from an entire GTM machine on a fixed date — and you do it three times, in genuinely divergent directions. The three PRFAQs become a negotiation, alignment and leadership instrument: they let you show stakeholders that you heard them, prove which paths don't add up, and then champion one recommendation with evidence rather than assertion.
Origin
The working-backwards / press-release-first method originates at Amazon (documented in the book 'Working Backwards', whose author Lenny had previously hosted). Rathi's additions are (a) treating the PRFAQ as a backwards plan for the whole launch machinery rather than just customer value, (b) the 'power of three' divergent variants, and (c) using the embedded quotes as forcing functions on other functions. He also ties it to Amazon's 'disagree and commit'.
Core principles
- 01You are not working backwards from a value prop; you are working backwards from a working GTM machine on a date.
- 02A press release is a one-pager: what we're building, by what date, why, the value prop, what the consumer says, what the business owner says.
- 03Three fully-thought-through alternatives beat one roadmap — comparison is a leadership tool.
- 04Recommend one. 'Here are three, you pick' is abdication.
- 05The FAQ section is where you force org-specific gates (compliance, legal, marketplace side-effects) to be answered.
- 06Once one is chosen: disagree and commit — leadership must stop relitigating.
How to run it
- 1
Fix the GTM date and write the one-pager backwards from it
Start from the launch date and the press release you want to exist on that day: what we built, the value proposition, why it's noteworthy, and what money and machinery had to exist for it to work.
Pro tip Because you start with a date, the PRFAQ doubles as a negotiation tool — take it to the VP Engineering and ask 'by this date, can we build all of this?'
- 2
Use the quotes as forcing functions on other teams
The customer quote is what you need marketing and pricing to make true. The business-owner quote is a goal commitment ('within three months we will have achieved this much'). Take each quote to its owner and get agreement.
Pro tip Use the quotes to find your alliances — and to surface, early, the people who are going to derail the launch.
Watch out If someone says they can't ship it, the date, the goals and the scope all move together. Don't quietly keep the old date.
- 3
Write three divergent variants, fully thought through
Not three flavours of one idea — genuinely divergent. E.g. for a membership launch: three tiers instead of two; or a tiered pricing programme with points instead of a membership; or the same offer aimed at a different segment. Each must be fully reasoned to a conclusion.
Pro tip Divergence is what lets you say 'I considered your point of view and I built the story around it — but unfortunately it doesn't add up, so we rejected it.'
Watch out Half-baked alternatives are worse than none — they look like strawmen and destroy trust with the leadership team.
- 4
Build organisation-specific FAQ gates
Encode the mandatory questions your company cannot ship without: at Jupiter (financial services) every PRFAQ must show compliance sign-off and legal review; at Swiggy (three-sided marketplace) it must answer implications for restaurant partners and each segment of delivery partners.
Pro tip The gates make PMs on one side of a marketplace consult the other sides — which surfaces weird, valuable correlations before launch.
- 5
Recommend one, then commit
Choose one option on the company's behalf based on what you know about market, company and strategy. If you're 80% right, pull elements of variants 2 and 3 into it. Once aligned, everyone commits — disagree and commit.
Pro tip This also works upward: when a CXO says 'let's build this', reply with three ways to build it plus the reason you're not building theirs.
Watch out Presenting three options with no recommendation makes you a facilitator, not a product leader.
In the wild
Rathi's worked example: launching a membership programme. PRFAQ 1 does three tiers instead of two. PRFAQ 2 abandons membership and builds a tiered pricing programme with points instead. PRFAQ 3 keeps the format but aims at a different segment. All three are fully reasoned.
→ Leadership can compare and contrast rather than react to a single roadmap, and the PM's rejection of an exec's idea is backed by a written, complete story.
On Swiggy's three-way marketplace, a change to delivery-partner earnings-per-hour touches cost per delivery, which touches delivery fee. The PRFAQ explicitly demanded answers on the implications for restaurant partners and for each delivery-partner segment.
→ PMs on one side of the marketplace started consulting the others by default, producing crisper products that worked for all sides.
Common mistakes
Working backwards only from the customer value proposition
The customer question ('will they love it, will they pay for it, is it noteworthy?') is only one part. If you don't work backwards from the entire launch machinery on a date, you get a lovely narrative and an unshippable plan.
Presenting one roadmap after a wide discovery phase
You heard many stakeholders during discovery. Showing one plan makes every unincorporated contributor feel ignored — 'didn't they listen to my interesting point?' Three divergent PRFAQs demonstrate that their idea was built out and honestly rejected.
Is it for you?
Best for
Senior PMs and heads of product proposing a bet that needs cross-functional sign-off and executive alignment
Not ideal for
Small reversible changes, bug fixes and routine backlog work — three PRFAQs for a copy tweak is theatre
From the transcript
“I really truly believe in the power of three”
“so instead of one you find the other set of all the things that need to come together to for that press release to go…”
“this alternative I considered your point of view and I created a story but unfortunately it's not adding up so we rejected it so now…”
“it's important for the product manager to recommend one”
From the episode
The full-stack PM
Anuj Rathi (Swiggy, Jupiter Money, Flipkart)