पाठशाला Pathshala · विचार Vichār, The idea · Lesson 25 · Scale

Adjacent markets: sequencing the second act

Rank adjacent markets by shared customer, shared channel and shared capability before you enter any of them, then enter them one at a time.

Pathshala, The Founder Library · 11 October 2026 · 6 min read

A wide banyan tree with aerial roots spreading out around its trunk in sunlight. — Photograph: BD Jewel · Pexels

The first act of a company is finding one thing that works. The second act is deciding what comes next, and it is where a surprising number of good companies stall. The temptation is to pick the biggest market in sight. The evidence says to pick the closest one, and to measure closeness before size.

This lesson explains why most second acts fail, gives three things an adjacent market can share with the core, and turns them into a ranking you can do in an afternoon. Two cases show what sharing the right things looks like, one Indian and one global. The closing section is the quarterly review that sequences the moves after the first.

Why most second acts fail

Chris Zook and James Allen of Bain studied companies that grew by moving into markets next to their core and published the result in Harvard Business Review in 2003. Their headline was that adjacency growth fails three-quarters of the time. Their explanation was more useful than the number: expansion works only when it is built around a strong core business, the best source of adjacency opportunities is inside the company’s strongest customers, and the companies that do it well develop a repeatable formula, screening each move before they make it.

Zook came back to the subject in 2007 with the opposite failure. Some companies leave their core too early for a fashionable market; his example was Bausch & Lomb, which diversified beyond contact lenses and later sold the new businesses at a loss. Others cling to a declining core for too long, as Polaroid did. Both misjudged where the core stood. For a founder the practical reading is simple: a second act is a move away from the core, and the further it moves, the more of what made the first act work is left behind.

Three things an adjacency can share

The customer. The same person or company buys the new product. You already know them, they already trust you, and the cost of reaching them is close to zero. This is the cheapest thing to share and the one Zook and Allen point to first.

An Indian thali with rice and several small bowls of different dishes on one plate. — One diner, one plate, many dishes. The cheapest adjacency is the next bowl for a customer already at the table. Photograph: Ryshy S · Pexels

The channel. The same route reaches the buyer: the same app, the same sales team, the same distributors, the same marketplace listing. A shared channel means the new product can be sold without building a new go-to-market, which is usually the most expensive part of a new business.

The capability. The same thing you are good at delivers the new product: the same technology, the same operations, the same data, the same supply chain. A shared capability means you are likely to be better than a new entrant from the first day, rather than having to learn.

An adjacency that shares all three is barely a new market; it is a new product line. One that shares two is a real adjacency with a good chance. One that shares one is a stretch that needs a strong reason. One that shares none is a new company with the old company’s money, and the honest way to treat it is as a new company: its own team, its own number, its own case for funding.

Ranking the candidates

List every adjacent market the team has discussed, usually three to six. For each, score the three dimensions: two points if it shares the dimension fully, one if partly, none if not. Do it with evidence: the share of your current customers who would buy, from a survey or from sales calls; whether the current channel already carries comparable products; whether the capability exists today or would have to be built. Then sort by the shared score, and only within the same score by the size of the market.

Start the list where Zook and Allen said the best opportunities sit: inside the strongest customers. Take the twenty accounts or the thousand users who pay most and stay longest, and ask them what they buy next to your product, from whom, and what annoys them about it. Write down every answer that comes up more than three times. Those are adjacency candidates with the customer dimension already scored at two, because the people who would buy them are the people who already buy from you. Candidates that come from a market report or a competitor’s launch start with the customer dimension at zero until a conversation proves otherwise.

Ranking by what is shared before size feels wrong, because the large market is more exciting. It is right because the chance of winning an adjacency falls fast as the steps away from the core rise, and a large market entered at three steps away has a smaller expected value than a modest one entered at one step. The figure below scores three candidates. Try making the largest market share nothing, and watch where it lands.

Rank by what the next market shares with you before you rank by how big it is.

Two adjacencies that shared the right things

Quick commerce from food delivery. Swiggy built Instamart beside its food business, selling groceries to many of the same customers, reached through the same app and the same habit of ordering in. Its Q4 FY 2025 letter to shareholders reported Instamart gross order value of ₹4,670 crore for the quarter, up 101 per cent, against ₹7,347 crore for food delivery, up 17.6 per cent. The letter also notes that about 30 per cent of the users Instamart acquired in the previous six months were new to Swiggy altogether, which means the adjacency had begun to feed the core rather than only draw on it. Customer: shared. Channel: largely shared. Capability: partly, because running dark stores is a different operation from dispatching restaurant orders, and the same letter shows Instamart still losing ₹840 crore at the adjusted EBITDA level that quarter. Two and a half of three.

Cloud infrastructure from retail. Amazon Web Services shared almost nothing with Amazon’s retail customers or channel. What it shared was capability. In 2006 Jeff Barr reported Jeff Bezos describing the undifferentiated heavy lifting of running web applications, which Amazon called muck, and noting that developers routinely spent 70 per cent of their time on that backend work. Amazon had already built that muck for itself. By the 2015 letter to shareholders, AWS was reaching $10 billion in annual sales. One of three, and a deep one; it worked because the capability was rare and Amazon could afford to build a new customer base and channel around it. Most companies with one shared dimension cannot.

Sequencing: one step at a time

The order matters as much as the choice. Enter the adjacency that shares most, run it until it has its own repeatable sales and its own retention, and only then enter the next. Each successful adjacency changes the core: a company that has added a product for the same customer now has two products to sell, a larger customer relationship and often a new capability, which can make a market that was three steps away only two. That is how a sequence of near moves reaches a far market that a single leap would have missed.

Two warnings. Do not enter two adjacencies at once; the team that built the first act cannot build two second acts in parallel, and both will be starved. And do not let the adjacency eat the core: the shared customer is also the one who will leave if the core product slips while attention is elsewhere. The [wedge lesson](/library/wedge-small-enough-to-win) covers the expansion path that a company should write down before it needs a second act; this lesson is what to do when the path forks.

The adjacency review, once a quarter

Keep one table: every candidate adjacency, its score on customer, channel and capability with the evidence for each, the size of the market, and its rank. Review it once a quarter. Re-score after every product launch, because the core has changed. Promote a candidate to a live project only when the one before it has repeatable revenue and the core’s own retention has not dipped. For a live adjacency, report three numbers to the board each quarter: revenue, the share of it from existing customers, and the core’s retention. If the second falls and the third falls together, the adjacency is drawing on the core rather than adding to it, and the team should decide whether to fix it or stop.

Figures were checked in October 2026 against the sources below. The scoring is this library’s method. Nothing here is investment advice.