Sun Tzu for startups: competing without imitating a large company

A startup does not need to replicate a corporation's entire catalogue to put forward a different offer. The key is to find a bounded problem it can solve with its resources and to check that this difference matters to the people who would buy the product.

Applying Sun Tzu to a startup requires setting aside the idea that small means agile and large means slow. The guide to Sun Tzu applied to business uses analogies, not rules about company size. The following case is fictional and makes it possible to examine a product choice, its tests and its limits.

What it means to focus against a larger competitor

The textual link is in emptiness and fullness in chapter 6: your own concentration, the opponent's disposition and adaptation. Its business version asks what need a team can meet without trying to match all of another company's capabilities at the same time. That formulation is a contemporary interpretation, not a business instruction written by Sun Tzu.

An underserved space is not the same as a proven opportunity. A product may be absent because the problem is rare, because it is hard to charge for solving it, or because users prefer another procedure. The absence of a visible product leaves an open question; it does not confirm that demand exists.

Weakness must be expressed in relation to a task. A platform can offer many functions and still be poorly suited to a specific process. That does not make its provider a weak company, nor does it mean it is unaware of that process. It may have chosen not to prioritise it, or it may solve it through another configuration.

Sun Tzu's vocabulary in business helps keep those limits in view. 'Concentrating forces' here means defining work and commitments. It does not imply harassing the competitor, deceiving its customers, or assuming it will disappear once an alternative is launched.

A fictional case: simplifying maintenance reports

Imagine a startup of four people: three build the product and one combines implementation and customer support. Its possible market is small maintenance companies whose technicians log visits from a mobile phone. The initial goal is to see whether it can reduce duplicate data entry between the technician's report and the administrative review.

In the example, a corporation offers a broad platform that includes planning, inventory and invoicing. The startup does not have the resources to replicate that catalogue. Nor has it shown that users want to replace it. Its hypothesis is narrower: some companies need to improve report closure without changing the rest of their tools.

The segment is defined by that process, not just by the size of the customer. A small company with highly customised jobs may need something different from another that carries out repetitive interventions. Mixing them in the same test would make it harder to understand which need is being validated.

Suppose the team sets aside six weeks for a prototype. That timeframe belongs to the fictional case; it is not a universal recommendation. The scope includes recording an intervention, reviewing the data and exporting the result. It leaves out full accounting, warehouse management and specific adjustments for each possible buyer.

The proposed differentiation is to complete that journey with fewer manual transcriptions. 'Simpler' would be an imprecise promise if it were not clear which task is being simplified. By contrast, seeing whether the technician and the administrative team can close the same report without re-entering information allows a concrete difference to be discussed.

The test should not be limited to asking whether the app is liked. Within the example, the point is to observe the current process, carry out an equivalent task with the prototype and note where help is needed. It also matters to check whether the person using the product and the person approving the purchase value the same outcome.

Errors and support time are part of the assessment. If each implementation requires several days of bespoke work, the team may be building a custom service rather than the product it imagined. That discovery is not necessarily a failure, but it changes the resources needed and the promises it can support.

Relying on a single person to implement and handle incidents introduces another limitation. Getting more trials than that person can support would not prove growth capacity. The case calls for examining the work after the sale, not just how many people agree to a demonstration.

The test can lead to different outcomes. It may confirm a repeatable improvement, show that the current tool already solves the problem once configured, or reveal that the saving does not justify the change. The broader investigation of those possibilities belongs to market niche hunting; here, what matters is how they shape this team's decision.

When specialisation stops making a difference

The corporation could add a similar feature, change its offer or maintain its priorities. None of those responses should be taken for granted. Building the whole project on the assumption that the competitor will not react confuses another party's decision with a condition under your own control.

Nor is it enough to observe a similar feature and conclude that the difference has disappeared. You would need to compare the journey users need, the implementation and the support. The criterion remains the problem being addressed, not the number of matching boxes in two catalogues.

There are signs that would require the case to be reviewed: every customer asks for a different product, integration prevents the result from being used, or no one considers a change of process necessary. Adding more features to avoid recognising those signs can undo the initial focus without solving its cause.

Economic pressure does not turn a hypothesis into an opportunity either. Reading Sun Tzu in a business crisis distinguishes coordination and urgency from a desperate bet. Needing revenue does not prove that a new product line will produce it in time.

The outcome of the example is left open. A reasonable decision could be to continue with a limited scope, reframe the proposition, or stop the test. What makes it possible to assess it is the relationship between observed need, delivery capability and the evidence obtained. The size of the company does not answer any of those questions on its own.

Before expanding the product, check whether the observed difference repeats without depending on exceptional attention. An isolated case does not prove that the team can sustain the same promise for other customers.

View all articles in Sun Tzu in Business: Resources, Competition, and Decisions