पाठशाला Pathshala · विचार Vichār, The idea · Lesson 21 · Build

Pick a wedge small enough to win and large enough to matter

Choose the narrow first use case that gets you in the door and the expansion path that justifies it. Two tests, both with numbers: can this team lead the segment, and is leading it worth a company?

Pathshala, The Founder Library · 11 October 2026 · 7 min read

Three metal chisels laid side by side on a wooden surface, seen from above. — Photograph: Collab Media · Pexels

Paul Graham’s advice on startup ideas has a picture at its centre. You can dig a hole that is broad but shallow, or one that is narrow and deep, like a well. The well is where companies start. The hard part, which the picture leaves to you, is that the well has to be narrow enough to dig with the people you have and deep enough to reach water.

This lesson turns that into two tests with numbers. The first says whether the wedge is small enough for this team to win. The second says whether winning it, and walking the path that follows, is large enough to matter. A worked example runs both, the figure lets you run your own, and the closing section is the monthly check that keeps the wedge honest once you are inside it.

Two tests a wedge has to pass

A wedge is the narrow first use case that gets a company in the door: one job, for one countable set of customers, done so well that they recommend you to each other. Facebook’s first site was exclusively for Harvard students, and even after it opened up, it remained for students at specific colleges for quite a while. Graham’s description of why is the most useful sentence on the subject: sometimes the right unscalable trick is to focus on a deliberately narrow market, like keeping a fire contained at first to get it really hot before adding more logs. Peter Thiel makes the same argument from the other end in Zero to One: start with a small market you can dominate, then expand to adjacent ones.

Both halves matter and founders usually check only one. A team in love with a large market picks a wedge too wide to lead, spends two years as one of eight vendors nobody recommends, and never gets the word of mouth that a leader gets for free. A team in love with focus picks a wedge so narrow that leading it produces a pleasant small business and no path out. The tests below are designed to catch both.

Small enough to win: the arithmetic of leading a segment

Leading a segment means holding enough of it that customers name you when they talk to each other. A working threshold is 40 per cent of the accounts; the exact number matters less than having one. Count the accounts in the wedge from a real list, district by district or industry directory by directory, not from a market report. Then count how many accounts the team can actually sign in a month today, from the last three months of closes rather than from the plan. Divide 40 per cent of the wedge by the monthly rate and you have the months to lead it.

Two years is a sensible ceiling. Beyond it the money runs out before the word of mouth arrives, and a competitor with a narrower wedge inside yours will have led its slice first. If the arithmetic says five years, the wedge is too wide for this team: cut it by geography, by size of customer, by the one job you do best, until the months fall under twenty-four. A wedge you cannot lead is not a wedge. It is a market you are participating in.

There are four ways to cut a wedge narrower, and the best wedges use two of them at once. Geography: one state or one cluster, where a sales visit costs a bus ticket and customers meet each other at the same association dinner. Size: one band of customer, because a five-person firm and a fifty-person firm buy differently even when they do the same work. Job: one job done end to end, rather than five jobs done partly. Trigger: customers at one moment, such as those switching from a spreadsheet this year because a new rule or a new competitor forced it. Each cut lowers the count of accounts and raises the close rate, because the product, the pitch and the references all fit more tightly. Keep cutting until the months to lead fall under twenty-four.

Large enough to matter: the path, written down

The second test asks what winning the wedge is worth. Revenue from leading it is the 40 per cent of accounts times what each pays a year. That number is almost always too small on its own, and that is fine, because the wedge is the door and not the house. What matters is the path: the next products sold to the same accounts, and the next segments that look enough like the first that the product and the sales motion carry over. Write it down as a sequence of steps, each with its own count of accounts and price, and add them up. The ratio of the whole path to the wedge is the expansion multiple.

A multiple of three to five is a path most teams can walk with evidence. A multiple of twenty needs a sentence for every step explaining why the customer, channel or capability carries over; the [adjacent markets lesson](/library/adjacent-markets-sequencing-second-act) gives the ranking method. If the whole path still adds up to less than a venture-scale company, say ₹100 crore of annual revenue for a team raising venture money, the wedge may be right and the funding wrong, which is the subject of a different lesson.

A worked example: diagnostic labs in Maharashtra

A team in Nashik builds software for independent pathology labs: sample tracking, reports delivered to the patient’s phone, billing to referring doctors. Their list of independent labs with two to ten staff across the tier-2 cities of Maharashtra has 1,500 names. They charge ₹60,000 a year. Over the last quarter they have signed 25 labs a month. Leading the wedge means 600 labs, which at 25 a month is 24 months: small enough to win, just. Revenue from leading it is ₹3.6 crore a year.

A row of sample tubes with orange caps standing in a laboratory rack. — Fifteen hundred labs you can name is a wedge. ‘Healthcare’ is not. Photograph: Charlie-Helen Robinson · Pexels

Now the path. The same product sold to similar labs in Karnataka, Telangana and Gujarat triples the accounts. Payments and referral-doctor settlement add a second product to the same labs. Reagent and consumables procurement through the platform adds a third with a larger ticket. Written down with counts and prices the team reaches a multiple of about twelve, or ₹43 crore. That passes the first test and fails the second. The honest response is not to widen the wedge. It is to find the step that changes the multiple: in this case, home sample collection for the labs, a logistics business with a much larger revenue per account, which takes the multiple near thirty and the path past ₹100 crore. Put your own numbers in.

Narrow enough that this team can lead it in two years. Deep enough that leading it opens a door worth walking through.

Wedges that fail, and how to tell early

The wedge defined by a word, not a list. “SMBs”, “Bharat”, “mid-market D2C”. If you cannot count the accounts from a list you could print, you cannot do the first test, and you are not yet in a wedge. The [positioning lesson](/library/positioning-sentence-that-decides-who-buys) is the place to start.

The wedge that is a feature of the expansion product. If the first thing you sell is something the eventual platform would give away, an incumbent with that platform will give it away. A wedge needs to be valuable on its own for the customer who buys it, even if the path never happens.

The path that changes customer. Step two sells to the lab owner; step three sells to the hospital chain. Each change of buyer is a new company with a new sales motion. A path is credible when most of its steps sell to the customer you already have or to one who looks just like them.

The wedge you have already outgrown. Leading the segment and then staying in it for another two years because it is comfortable. The signal is the close rate falling in the wedge while inbound arrives from outside it. That is the moment for step two.

A monthly check on the wedge

On the first working day of each month write four numbers on one line: accounts in the wedge, accounts signed, share of the wedge you hold, and the trailing three-month close rate. Divide the gap to 40 per cent by the close rate and you have the months to lead, which should fall every month; if it rises two months running, the wedge is too wide or the product is not yet good enough, and the team should decide which. Once a quarter, rewrite the path with the evidence you now have: the step you are surest of, the step you are least sure of, and the multiple. When the share passes 40 per cent, take the first step on the path, and start the same check for the next segment.

The worked example is illustrative; the labs, counts and prices are not drawn from any company. Nothing here is investment advice.