The business should carry out research, but should spend far less than it is considering and choose its methods to match the budget.
The case against spending is real. Money spent on research is money not spent on stock, equipment or promotion, and for a business with very limited funds that trade-off is severe. Research also delays the launch, during which a competitor may reach the market first. Most seriously, a small budget buys a small sample, and a small sample produces findings the business cannot rely on — so it may spend a significant part of its funds and still not know whether the product will sell. Spending scarce money on unreliable information is the worst of both outcomes.
The case for spending is stronger. Launching without research risks far more than the research would cost: producing stock nobody wants, pricing the product where nobody will buy, or aiming it at the wrong customers. For a business with very limited funds a failed launch may not be survivable, so the relevant comparison is not the cost of research against zero but against the cost of being wrong. Research also improves the launch even when demand is confirmed, by indicating price and where to sell.
The decisive point is that this is not a choice between expensive research and none. Secondary sources — government statistics, market reports, competitors' websites and prices — cost almost nothing and would tell the business the size of the market and what rivals charge. Cheap primary research, such as observing shoppers or asking a modest but properly chosen sample, adds what secondary sources cannot. Together these cost a fraction of what the business is contemplating while removing most of the uncertainty that matters.
I therefore reject spending a significant proportion of limited funds, not because research is unnecessary but because that amount is not needed to obtain the information required. The business should begin with secondary sources, add targeted primary research where a real question remains, and keep the remainder for the launch itself. If a genuine uncertainty survives all of that — and it is one large enough to threaten the business — then spending more would be justified.