Founders and new sales heads often say a sentence that sounds like ambition: “Our product is useful for any company with more than fifty employees, so our market is huge.” Narrowing it down feels like turning away revenue. A salesperson told to focus on one kind of customer may feel the same, wondering why a willing buyer from a different industry is treated as a distraction.
The belief has a grain of truth, which is why it survives. A product may indeed be usable by many companies, and a wide funnel does produce more conversations. But “could use it” and “is a good customer for us” are different statements, and the distance between them is where money is lost.
What an ICP actually is
An ideal customer profile, or ICP, describes the kind of company that gets the most value from your product and gives you the most value back. It is about the organisation: industry, size, geography, technology in use, how it buys, the problem it has. It is not the same as a buyer persona, which describes the person inside that company, such as the head of finance operations, with her goals, worries and how she judges vendors. You need both, and the ICP comes first: you decide which companies to approach, then which people inside them to speak to.
What the broad market actually costs
Take two invented segments from a company that sells collections software. Segment A is mid-sized manufacturers with several plants. Segment B is small trading firms with a single office. Both have a budget. Both can use the product.
From the company’s records: in A, one opportunity in four is won (25%), the average first-year contract is ₹8 lakh, and 8% of customers leave each year. In B, 12% are won, the average contract is ₹5 lakh, and 22% leave each year.
Out of 100 opportunities, A produces 25 wins and ₹200 lakh of first-year contract value, while B produces 12 wins and ₹60 lakh. A is more than three times better before even looking at retention. Now add churn. A rough lifetime is one divided by the annual churn rate: about 12.5 years for A and 4.5 for B. Multiplying first-year value by that lifetime, which is crude and ignores growth, discounting and the fact that nobody can predict twelve years ahead, gives about ₹2,500 lakh against ₹273 lakh. The exact figures are not the lesson. The ranking is, and the gap is large enough that no plausible correction closes it.
Meanwhile both segments cost the same to chase. A rep spends a similar number of hours on a Segment B opportunity as on a Segment A one. The broad strategy does not just earn less per deal; it uses the scarcest resource, selling time, on the worst returns.
Other costs that do not show in a funnel
- Product drift. Customers outside your core keep requesting features that suit them, and the product team follows the loudest voice instead of the largest segment.
- Messaging that says nothing. When you speak to everyone, you cannot say “we understand multi-plant billing”, so your pitch becomes generic and is easily replaced by a cheaper competitor’s.
- Support load. Customers who do not fit often need more help and give less in return.
- Weak references. Eight customers in eight industries make a poor case study set. Eight in one industry give prospects a peer to call.
Building one from evidence
Start with your own customers, not with a wish. Take the accounts that bought fast, paid on time, expanded and stayed, and look for what they share: industry, headcount, revenue range, systems they already run, the event that made them buy, the job title of their champion. Then look at the worst accounts, the ones that churned or consumed support, and ask what they had in common.
Turn the pattern into a handful of criteria that a rep can check in five minutes from public information: perhaps “manufacturing, 200 to 2,000 employees, at least three locations, uses an ERP”. Add one or two signals that show a company is in-market. Score accounts against it, and let that score guide where outbound effort and marketing money go first.
What the myth gets right
A narrow ICP is not a rule against ever selling outside it. Early on, with very few customers, you do not yet have the data to know your best segment, and a deliberately wide first year is a way to learn. Also, a narrow definition can be wrong: if it is built from twelve customers, one unusual account can mislead it. And a segment may be small; if the best-fit market has only 400 companies in India, focusing on it may cap growth and force a later widening.
So the sensible position is not “turn everyone else away”. It is to know which customers you are optimising for, take outside ones deliberately and with open eyes, and review the profile every year as the product and the market change.
Figures and names in the examples are illustrative.
Leave a Reply