पाठशाला Pathshala · उत्पाद Utpād, The product · Lesson 30 · Scale
The product strategy document for a hundred-person company
At a hundred people the founder can no longer be in every product decision. A short strategy document that connects company goals to a few themes lets squads decide well without asking.
Pathshala, The Founder Library · 11 October 2026 · 7 min read
At thirty people the product strategy lives in the founder’s head and everyone can ask. At a hundred, eight squads are making decisions every day in rooms the founder is not in. Either they share a written strategy or they share a queue outside the founder’s door.
This lesson sets out what a product strategy is and is not, the sections of a document a hundred-person company can run on, how to give squads the right to decide, a worked example from Gurugram, and a quarterly routine that keeps the document alive.
What changes at a hundred people
Somewhere between fifty and a hundred and fifty people a product company reorganises into small teams, each owning a part of the product end to end. Henrik Kniberg and Anders Ivarsson’s 2012 description of Spotify’s structure gave the vocabulary many companies now use: a squad is “designed to feel like a mini-startup”, and ideally “each squad is fully autonomous”; squads working on related areas form a tribe, kept “smaller than 100 people or so.” The structure works only if autonomous teams pull in the same direction. Autonomy without a shared strategy produces eight good products that do not add up to one company.
The symptoms of a missing strategy are familiar. Squads bring the same trade-off to the founder every week. Two squads build overlapping features for different segments. The roadmap is a list of requests from the largest customers. Each squad’s metrics improve while the company’s do not. All of these are the cost of a strategy that has not been written down.
Strategy is a diagnosis and a choice
Richard Rumelt’s kernel of good strategy, from his book Good Strategy Bad Strategy, has three parts. A diagnosis: “an explanation of the nature of the challenge.” A guiding policy: “an overall approach chosen to cope with or overcome the obstacles identified in the diagnosis.” And coherent actions: steps coordinated to carry out the policy. His signs of bad strategy are just as useful: fluff, failure to face the problem, and mistaking goals for strategy. A target of doubling revenue is a goal. It becomes a strategy only when the document says what is stopping the company and how it will get past it.
Marty Cagan of SVPG puts the product version in Product Strategy Overview: product strategy decides which problems to solve, product discovery finds solutions that work, and delivery builds them. Its core is focus, choosing the few things that matter most, built on insights from data and customers, turned into action by giving teams objectives with context. The document described below is that idea made concrete enough for a squad lead to use on a Tuesday afternoon.
The document, section by section
1. Company goals. Three to five numbers for the year, copied from the operating plan: revenue, net revenue retention, gross margin, perhaps a customer count in a target segment. The product strategy serves these; it does not invent its own. 2. Diagnosis. One page on the one or two problems standing between the company and those numbers, with evidence: cohort data, win-loss notes, customer interviews. 3. Guiding policy. A paragraph that makes a choice, such as which segment the product is built for this year and what kind of advantage it will rely on. A good policy rules out options someone in the company wants.
4. Themes. Three to five problem areas the product will work on, each with the metric it should move and the squads that own it. A theme is a problem, not a feature: “finance teams cannot reconcile travel invoices” rather than “build reconciliation.” 5. Not doing. A list of things the company has decided against this year, with one line of reason each. This is the section squads use most. 6. Decision rights. What squads may decide alone, what they must consult on and what comes to the product head. 7. Review. The dates the strategy is read against results, and who is in the room. Six pages is enough. A strategy nobody finishes reading does not steer anyone.
A goal says where the company wants to be. A strategy says what is in the way and what the company will stop doing to get past it.
Write the diagnosis first and the themes last. Most teams do the reverse: they list what they already plan to build, group it into themes and write a diagnosis that justifies the grouping. The test of an honest diagnosis is that it surprises someone in the leadership team and kills at least one project that was already under way. Circulate the draft to squad leads before it is final and ask one question of each: which decision you face next quarter would this document settle? If the answer is none, the document is too abstract.
Letting squads decide without you
The point of the document is fewer escalations. Jeff Bezos’s 2016 letter to shareholders gives the principle: many decisions are “reversible, two-way doors” and can use a light process, and “most decisions should probably be made with somewhere around 70% of the information you wish you had.” The same letter offers “disagree and commit” for the cases where a leader has doubts but the team has conviction.
Turn that into rules in section six. A squad decides alone how to solve a problem inside its theme, what to build first and when to ship, provided the decision can be reversed within a release. It consults another squad when a change touches that squad’s metric or code. It brings to the product head anything that changes pricing, enters a segment the strategy excluded, removes a feature customers pay for, or cannot be undone. Write the rules down and then honour them. A founder who reverses squad decisions that sat inside the rules teaches every squad to ask first, and the document stops working. Pair this with a [writing culture](/library/writing-culture-decisions-in-documents) so decisions taken alone are still visible to everyone.
A worked example: corporate travel software in Gurugram
A Gurugram company sells corporate travel booking to Indian businesses. It has 104 people, eight product squads and annual revenue of ₹45 crore. The operating plan sets three goals: ₹70 crore of revenue, net revenue retention of 115 per cent, and gross margin held above 60 per cent. The founder spends two days a week settling product questions.
The diagnosis, drawn from churn interviews and cohorts, is that accounts stay while the travel manager who chose the product stays, and leave when she does, because the product is valued by one person and not by the finance team that pays. Enterprise deals stall on approval policies the product cannot express. The guiding policy: build for the finance team at companies of 200 to 2,000 employees, and make the product the place where travel spend is reconciled, not only booked. Three themes follow. GST invoice reconciliation, measured by the share of invoices matched automatically for input tax credit. Policy and approvals, measured by the win rate on deals over 500 employees. Self-serve onboarding for smaller companies, measured by days from sign-up to first booking. The not-doing list names leisure travel, a rebuilt mobile app and international expansion, each with its reason. Within a quarter the questions reaching the founder fall to a handful a week, nearly all of them about the items the document said should come to her.
Five ways the document fails
Goals dressed as strategy: a page of targets with no diagnosis. Too many themes: eight themes for eight squads means no choice was made. A roadmap in disguise: a list of features with dates, which squads will execute without judgement. No not-doing list: every request remains possible, so every request is negotiated. Written once: a document from January that nobody has opened since March is a souvenir. Each of these is visible in the first review, which is why the review is part of the document.
The quarterly strategy refresh
Rewrite the strategy once a year with the operating plan, and refresh it every quarter in a half-day meeting of the founder, the product head and the squad leads. Read each theme’s metric against the target. Ask whether the diagnosis still holds, using the quarter’s cohorts and win-loss notes. Add to or confirm the not-doing list. Count the decisions that came to the founder and check how many were outside the rules in section six; each one inside the rules is a sign the document is unclear there. Then let each squad set its quarter’s objectives under its theme, in the format of the [roadmap as a set of bets](/library/roadmap-as-set-of-bets), and publish the refreshed document the same week.
The Gurugram company is illustrative. Quotations are from the sources below, checked 11 October 2026.
Sources
- Richard Rumelt, The perils of bad strategy, McKinsey Quarterly, June 2011 — The kernel of good strategy: diagnosis, guiding policy, coherent actions; drawn from his book Good Strategy Bad Strategy.
- Marty Cagan, Product Strategy Overview, SVPG, February 2020 — Focus, insights, action and management.
- Henrik Kniberg and Anders Ivarsson, Scaling Agile @ Spotify with Tribes, Squads, Chapters and Guilds, October 2012 — Squads as autonomous mini-startups; tribes under about 100 people.
- Jeff Bezos, 2016 Letter to Shareholders, Amazon — Two-way doors, deciding with about 70 per cent of the information, disagree and commit.