apply-probe-strategy

Apply Probe Strategy

Design a low-cost, deliberately limited action that reveals hidden opposition, competitive response, or customer reaction before committing full resources to a major move.

Why This Is Best Practice

Origin: Stratagem #13 of the Thirty-Six Stratagems: "Beat the grass to startle the snake" (打草驚蛇). The image is a traveller who beats the long grass before stepping through it — the snakes reveal themselves through movement before they can strike. In military strategy: conduct a probe before the main assault to reveal enemy positions, strength, and tactics. The cost of the probe is small; the cost of an ambush is potentially the entire campaign.

Adopted by: The probe strategy underlies lean startup methodology (minimum viable product as probe), regulatory pre-consultation processes (test regulatory response before committing to a product architecture), and iterative product development. Amazon's Press Release/FAQ process — writing the press release for a product before building it — is a probe: does this product sound compelling enough to build? A/B testing in product development is probe strategy at the micro-scale. Before Google launched Gmail to the public in 2004, it ran a private beta that revealed demand far beyond initial estimates and allowed scaling before public launch.

Impact: Strategic commitment without probing exposes full resources to hidden risks. A competitor you thought was weak may be preparing a response. A regulatory body you assumed was neutral may be developing policy. A customer segment you assumed wanted the product may respond to pricing in ways you did not predict. Probing converts unknown risks into visible, addressable ones before the point of irreversible commitment.

Why best: The probe's value is asymmetric: the cost of a good probe is small relative to the cost of a failed full commitment. The information value of the probe's result — whether it reveals a snake or confirms clear ground — is always greater than zero. Probing is not timidity; it is the application of information-gathering before investment, which is the fundamental discipline of evidence-based strategy.

Sources: Thirty-Six Stratagems #13 (Sawyer trans. 1994); Ries, The Lean Startup (2011); Klein, Sources of Power (1998) — naturalistic decision-making under uncertainty

Steps

Step 1: Identify the specific uncertainty the probe must resolve

A probe designed to reveal everything reveals nothing useful. Define precisely what you do not know and what you need to know before committing:

Uncertainty type	Probe design
Customer demand	Smoke test, landing page with "notify me," pre-sales without product
Competitive response	Announce in trade press; watch for counter-move
Regulatory posture	Pre-consultation meeting with regulator; informal briefing to policy staff
Channel acceptance	Pilot with one reseller or distributor; expand only after performance data
Pricing sensitivity	A/B test pricing on new cohort; shadow pricing before full rollout
Partnership interest	Non-binding LOI or exploration meeting before contract negotiation

One probe per critical uncertainty. Multiple probes in parallel work if the uncertainties are independent.

Step 2: Design the probe to provoke genuine response

A probe that is too hypothetical produces hypothetical responses. The probe must be real enough that:

• A customer who sees a smoke test ad responds because they actually want to buy, not because they're being polite
• A competitor who sees a market announcement responds with a real counter-move, not a PR statement
• A regulator who receives a pre-consultation briefing responds with substantive feedback, not a form letter

Indicators of a probe that is too hypothetical:

• Customer research question: "Would you buy this if it existed?" (always yes; no purchasing commitment)
• Competitor probe: "What do you think our competitor would do?" (speculation; not observed behaviour)
• Regulatory probe: Asking external lawyers rather than the regulator directly

Step 3: Execute the probe and observe — including what does NOT react

The probe's output has two parts:

1. What reacted: The competitor's counter-announcement, the customer segment's pre-sign-up rate, the regulator's letter
2. What did not react: Competitors who did not respond may be unaware, dismissive, or preparing a delayed response — all meaningful signals

Absence of reaction is as informative as reaction. If you priced a product publicly and the dominant competitor made no move, assess: (a) they did not notice; (b) they noticed but considered you non-threatening; (c) they are preparing a delayed response. Each implies a different next move.

Step 4: Interpret what the probe reveals before committing

Before moving from probe to full commitment, translate the probe's output into a decision-relevant insight:

• Probe confirmed the assumption: Proceed with the main move, adjusted for anything the probe revealed about execution details
• Probe revealed partial risk: Redesign the move to address the specific risk the probe revealed; do not proceed with the original plan unchanged
• Probe revealed a disqualifying risk: Abandon or substantially rethink the move; the cost of proceeding now is greater than the cost of redesigning
• Probe was inconclusive: The probe did not reveal the needed information; redesign the probe and run again before committing

The probe result must change something about the main move — otherwise the probe was not actually decision-relevant.

Step 5: Move to full commitment only after the probe resolves the uncertainty

The probe's purpose is to reduce the uncertainty of the commitment decision, not to delay commitment indefinitely. Once the probe has resolved the uncertainty it was designed to address:

• Commit fully if the probe was positive
• Adjust and commit if the probe revealed manageable risks
• Do not run additional probes on uncertainties the first probe already resolved — that is delay, not discipline

Rules

• One critical uncertainty per probe. Probes designed to answer too many questions produce unclear results. Isolate the highest-stakes uncertainty; design the probe for that question; run it; then address secondary uncertainties.
• The probe must be real enough to provoke genuine response. A probe that cannot provoke a real response cannot reveal real information. If you cannot design a real enough probe, the probe approach is not the right tool — consider scenario planning or expert interviews instead.
• Absence of response is signal. Do not interpret no response as confirmation. Analyse why the snake did not move: unaware, dismissive, or preparing.
• Act on the probe result. Running a probe and ignoring its output because it is inconvenient is worse than not probing. The probe's value is in the decision it enables, not the data it generates.
• Probing is not permission to delay indefinitely. Some decisions cannot wait for probe results. Know in advance the maximum time you will spend on the probe phase before committing regardless.

Examples

Product smoke test before building: A startup wants to know whether there is sufficient demand for an enterprise document management product before committing 6 months of engineering. Probe: build a landing page describing the product, run targeted ads to the intended buyer persona (VP Operations at companies with 100–500 employees), and measure sign-up rate for a "request early access" CTA. No product exists; the landing page is the probe. If sign-up rates exceed 3% from the ad cohort (industry benchmark for enterprise demand), the demand signal is positive. If 0.3%, either the targeting is wrong or the demand is thin. The probe costs $5,000 in ad spend and 2 weeks; the product would cost 6 months and $400K. The probe resolves the demand uncertainty before the commitment decision.

Regulatory pre-consultation for a fintech product: A fintech company wants to launch a product that involves novel application of existing financial regulation. Uncertainty: will the regulator treat the product as a licensed activity requiring prior approval, or as an exempt activity? A surprise enforcement action post-launch could require product withdrawal. Probe: request an informal pre-consultation meeting with the regulatory body's policy team. Present the product concept and ask for initial views on classification. The regulator's response — even an informal "this looks like a licensed activity, you should file for authorisation" — is the probe's output. The probe costs 2 months of regulatory counsel time; proceeding without it risks a product recall and enforcement cost.

Competitive response probe before major announcement: An established SaaS company plans to enter an adjacent market currently served by a niche competitor. Uncertainty: will the competitor respond with aggressive pricing to defend their position, or will they concede the market and focus on their core? Probe: brief a single industry analyst on the entry plan, attributing it to "a major SaaS platform." Monitor the competitor's public communications and customer-facing messaging over the following 30 days. If the competitor publicly announces defensive pricing or new product capabilities, the response is aggressive; budget for a price war. If there is no response, the competitor may be unaware or may lack the resources to respond — proceed with the original plan.

Common Mistakes

Probing on the wrong uncertainty: Running a customer demand probe when the actual uncertainty is regulatory posture, or a competitive probe when the actual uncertainty is internal execution capacity. The probe reveals what it measures; identify the correct uncertainty first.

Designing an obviously hypothetical probe: "Would you buy this?" surveys do not reveal purchase intent; they reveal conversational agreeableness. Probes must require real commitment to produce real signal (time, contact details, a small payment, a stated preference with trade-offs).

Ignoring inconvenient probe results: The probe that reveals a disqualifying risk is the most valuable probe — it saves the cost of a failed commitment. Dismissing an inconvenient probe result by saying "that was an outlier" or "the conditions were unusual" wastes the probe's value.

Probing indefinitely: Sequential probing without a commitment gate converts probe discipline into delay. Before the probe, define the criteria under which you will commit regardless of the result (either the probe resolves the uncertainty, or you commit after a fixed time window).

Treating the probe result as definitive rather than directional: A probe that confirms demand, competitive response, or regulatory posture in one context may not hold in another. Probe results are directional signals that reduce uncertainty — they do not eliminate it. Proceed after a positive probe with confidence, not certainty.