How to define and expand your ideal customer profile

Executive overview

Most companies define their ICP based on who is easiest to sell to — lowest CAC or highest inbound volume. That produces customers who churn, not scale.

The right ICP is built on maximum LTV, enforced org-wide, and structured to expand deliberately over time. A red-yellow-green framework makes this operational.

The ICP that matters is not what you put on your website — it's which accounts your salespeople actually choose to pursue.

The red-yellow-green framework

• Green: perfect-fit accounts; all outbound, ads, and SDR effort targets only these
• Red: do not sell, ever — tear up the contract even if it's a million dollars
• Yellow: peripheral segments; sell only on inbound; use to run controlled ICP experiments
• Example green profile: US-based SMBs in tech or industrials with a full-time ops role
• Example red: Asia, enterprise (5,000+ employees), healthcare, no ops person on staff
• Example yellow: English-speaking UK companies in tech SMB who come inbound

Pothole 1 — defining ICP by who buys, not who succeeds

• Inbound demand signals interest, not fit
• Customers who came in easily may not have been built for; they churn faster
• Lowest-CAC customers are often lowest-LTV customers
• A stay-at-home freelancer is easy to close on HubSpot; a 20,000-person global company has massively higher LTV even after a high acquisition cost
• Route ICP by LTV, not CAC — retention and lifetime value reveal who your product actually works for

Pothole 2 — starting too broad

• VC pressure pushes founders to claim a massive TAM from day one
• Trying to sell to Asia, enterprise, healthcare, and SMB simultaneously produces chaos, not growth
• A tight ICP is the fastest path from zero to $5M to $20M ARR
• You don't need to prove a multi-billion dollar TAM in year one — you need to convince investors you're on a trajectory to get there
• Narrative to investors: show the expansion path, not the starting surface area

Running ICP expansion experiments

• Once at scale (e.g., 20 salespeople), dedicate the majority to the core green ICP
• Assign 2–3 reps as dedicated ICP experiment teams targeting yellow segments
• You need only one expansion experiment to work to fuel the next growth phase (e.g., $20M to $100M)
• Experiment results drive the decision to promote yellow to green or retire it to red

AI and dynamic ICP management

• AI enables ICP analysis across far more attributes than humans can track manually
• Continuous evaluation — hourly or daily — across macroeconomic shifts, product changes, competitive moves, intent signals
• Salespeople's time is a scarce resource; AI identifies the optimal accounts to pursue on any given day
• From a pipeline of 127,000 accounts, AI surfaces the 17 that will most predictably generate maximum revenue today
• Outcome: faster, more predictable growth from $20M to $100M+