पाठशाला Pathshala · ग्राहक Grāhak, The customer · Lesson 23 · Scale
Choosing the second segment without losing the first
The second segment is where many companies that found fit lose it. Price three costs before you expand: acquiring the new customer, closing the product gap, and the margin your first customers take away when they trade down.
Pathshala, The Founder Library · 11 October 2026 · 7 min read
The first segment is found by obsession. The second is usually chosen by a board meeting, a large inbound enquiry or a competitor’s press release, and it is where many companies that found fit start to lose it.
This lesson sets out when to expand, how to find the candidates, the three costs to estimate for each, a figure that turns them into two years of contribution, how to fence a new offer so the first customers stay, and the guardrails that tell you to stop.
When the second segment is the right question
Geoffrey Moore’s Crossing the Chasm, first published in 1991, argues for a segment-focused approach: win a single demanding use case in one market before widening. The second segment is where that discipline is most often lost, because the first one now looks easy from the inside.
Three tests should pass before expansion is on the agenda. The first segment is still growing month on month without the founders closing every deal. It retains: cohorts flatten, as the [retention lesson](/library/retention-curves-and-flattening-test) describes, and churn has an owner. There is a reason beyond growth for growth’s sake: the first segment is approaching its ceiling in your channels, or a second segment is pulling the product toward it with real purchase orders, not compliments. If any test fails, the second segment is a distraction from the first.
Find the adjacent candidates
An adjacent segment shares at least two of three things with the first: the buyer (the same role, so the pitch carries over), the channel (the same way of being found, so acquisition carries over) and the product (most of what they need already exists). A clinic software company has diagnostic labs and pharmacies as candidates; hospitals share the product but not the buyer or the channel, and are a different company.
Interview ten to fifteen prospects in each candidate before modelling anything, using the method in the [segmentation lesson](/library/segmentation-that-changes-what-you-build). Write down, for each candidate, the three must-have capabilities you do not have, the price they pay today for the alternative and how they would find you. That is the raw material for the three costs.
Look for pull before push. Count the inbound enquiries from each candidate in the last six months, the current customers who already belong to it in part (the clinic that also runs a small lab), and the deals lost because a capability was missing. A candidate that is already knocking on the door will cost less to acquire than the model assumes; one that nobody has asked about will usually cost more.
Three costs to estimate before you expand
Acquisition cost. A new segment usually means a new channel, a new message and a sales team learning a new objection. Estimate CAC from a small paid test or twenty founder-led calls, and divide it by monthly gross margin per customer to get payback. David Skok’s SaaS metrics guide puts the warning line at twelve months, beyond which profitability becomes anaemic, and notes that many of the best SaaS businesses recover CAC in five to seven.
Product gap. Turn the must-have list into engineering months and rupees, and add the months of distraction for the team serving the first segment. Founders underestimate this cost more than the other two, because the first segment’s roadmap does not stop while the new one is built.
Cannibalisation. A cheaper or simpler offer for the new segment will tempt some first-segment customers to trade down, and a price change made for expansion can drive some away altogether. Netflix lost about 800,000 US subscribers in the quarter after it split streaming and DVD rentals into separately priced plans in 2011, and told shareholders that many long-term members felt shocked by the change. Estimate cannibalisation as a share of the first segment’s monthly margin, from interviews with your own customers about the new offer.
Model two years of contribution
Put the three costs into one line: cumulative contribution over twenty-four months, starting below zero at the build cost, falling further as acquisition spend runs ahead of margin, and climbing as customers accumulate. Subtract cannibalised margin every month. Compare candidates on three numbers: CAC payback, the break-even month and the deepest point of the cash hole, which is what the expansion will actually cost you to fund.
Hold the deepest point against your runway. An expansion whose cash hole is ₹30 lakh is cheap for a company with ₹5 crore in the bank and fatal for one with ₹60 lakh and nine months to go; the [runway lesson](/library/runway-how-many-months-you-really-have) has the arithmetic. If the hole is larger than a quarter of the cash you hold, either slow the pace of acquisition, which shallows the hole and delays the break-even, or wait until the first segment funds it.
Slide cannibalisation from 6 per cent to 2 per cent at the starting values and the two-year result rises from about ₹26 lakh to about ₹45 lakh, with break-even four months earlier. No change in CAC or product does as much. The cheapest segment to acquire can be the most expensive to own if it eats the first one.
Fence the offer so the first segment stays
A fence is a difference between the offers that matters to the new segment and would cost the first segment something it values. Fences can be features (the lab plan has no prescription module), limits (one counter, one location, a cap on invoices), devices or channels (sold only through distributors, or only on mobile), or support (no named account manager). Netflix’s ₹199 mobile plan in India in July 2019 is a clean fence: one smartphone or tablet at a time in standard definition, a price a new segment could clear, and nothing a household watching on a television would trade down to.
Test the fence before launch by asking twenty current customers whether they would switch to the new offer and why. If more than a few say yes, the fence is in the wrong place. Never move existing customers onto worse terms to make the new offer look better; grandfather them and say so.
The cheapest segment to acquire can be the most expensive to own, if it eats the customers who made you.
A worked example: clinics to labs or pharmacies
A Pune company sells clinic software and earns ₹20 lakh a month of gross margin from clinics. It considers two segments. Diagnostic labs: CAC about ₹25,000, ₹4,000 of monthly margin per lab, ₹15 lakh of product work for sample tracking and report delivery, fifteen new labs a month and almost no overlap with clinics. The model shows payback in about six months, break-even in month 18 and about ₹35 lakh of contribution by month 24, after a cash hole of about ₹27 lakh.
Pharmacies: CAC about ₹15,000 through the distributors that already supply them, ₹2,500 of monthly margin, ₹6 lakh of product work and twenty-five new pharmacies a month. Unfenced, the pharmacy plan includes billing that small clinics would happily use at a lower price; interviews suggest 6 per cent of clinic margin would move. The model gives break-even in month 20 and about ₹26 lakh by month 24, less than the labs. With a fence, a cap of one counter and no appointment module, interviews suggest the loss falls to 2 per cent, break-even comes in month 16 and the two-year figure rises to about ₹45 lakh. The company chooses fenced pharmacies, and builds the lab product the following year.
The monthly expansion review
Once a month while the second segment is young: first-segment retention and growth, by cohort, against the months before expansion; a fall is the first sign the team’s attention has moved. Downgrades and switches from first-segment customers to the new offer, with a reason for each. CAC and payback in the new segment against the model. Product gap remaining, in engineering months. Cumulative contribution against the model line, and the gap explained.
Write the stop rule before you start, for example: if first-segment net revenue retention falls two quarters running, or the new segment’s payback is above twelve months after six months of selling, pause and review. A rule written in advance is easier to keep than one argued in the moment.
Give the second segment its own small team, its own metrics and its own weekly review, and leave the team that serves the first segment alone. When the same people sell and support both, the newer and more exciting segment wins every argument about time, and the first segment learns about it from slower replies. Separate the work until the second segment has its own repeatable sale, then merge what can be shared.
The worked example is illustrative; replace every figure with your own interviews and tests before relying on the model.
Sources
- Geoffrey Moore, Crossing the Chasm (Harper Business, 1991, 1999, 2014), author’s page — A segment-focused go-to-market approach that targets a single demanding use case first.
- David Skok, SaaS Metrics 2.0: A Guide to Measuring and Improving What Matters, For Entrepreneurs — Profitability becomes anaemic beyond 12 months to recover CAC; many of the best recover it in 5–7 months.
- TV Technology, Netflix reports loss of 800,000 domestic subscribers in Q3, 26 October 2011 — Attributed largely to separating streaming and DVD plans at $7.99 each.
- Netflix, Netflix launches mobile plan for India, 24 July 2019 — ₹199 a month, one smartphone or tablet at a time, standard definition.