Rocksalt.ai moved beyond a simple persona ("CEO") to a behavioral ICP. Their ideal customer is a CEO who is already trying to post on LinkedIn 1-2 times a month and has 2k-10k followers. This sharp, behavior-based definition allows them to instantly identify high-propensity buyers before a call even begins.

Firmographics like industry or company size are often just proxies, not the core reason for purchase. The strongest signal for a successful customer is their specific use case. Capturing this data in the CRM is critical for building an accurate and effective ICP.

When selling to what you believe is a single ICP, but some buyers have intense 5/5 "Pull" and others have a mild 2/5, your ICP definition is flawed. The difference in their behavior is the key signal. You must diagnose the non-obvious differences between these groups to define your true, high-intensity ICP.

Founders often believe their ICP is a theoretical construct for their website and pitch decks. In reality, a company's true ICP is determined by the customers the sales team is actively pursuing and successfully closing, which can reveal a critical disconnect from the intended strategy.

A traditional ICP mixes high- and low-intent buyers, yielding mediocre 20-30% close rates. An ICP based on "pull" focuses exclusively on the specific situations that create urgent, blocked demand. This forces hyper-specificity and builds a more efficient GTM motion by targeting a cohort with a near-100% close rate.

Stop defining your Ideal Customer Profile with abstract firmographics. Instead, feed context from your best closed-won deals into an AI and ask it to find public data that signaled their specific pain *before* they engaged you. This reverse-engineers a truly effective, data-driven targeting model.

Ditch the aspirational "Ideal Client Profile," which represents a rare, perfect-world scenario. Instead, build a "Target Client Profile" that defines which customers will perceive the most meaningful value from your offering. This provides a realistic, operational benchmark for qualifying leads.

Don't just target the same job titles as your best customers. Dig deeper into the buyer's professional history (e.g., a COO with a 20-year sales background). This backstory is often the true indicator of an ideal fit, allowing for more precise and effective targeting.

Defining an ICP based on who you *want* to sell to is flawed. A "Pull"-based ICP is defined reactively: it's the specific group of people currently experiencing such an urgent, blocked project that it would be illogical for them *not* to buy your solution right now.