Ideal Customer Profile (ICP) Skill

Ideal Customer Profile (ICP)

The sharp definition of who you serve — used by marketing for targeting, sales for qualification, and product for prioritization.

When to use this skill

Defining ICP for a new product / segment
Refining ICP from observed closed customers
Auditing when pipeline quality / conversion is poor
Pre-fundraise to articulate market position
Sales qualification rubric build
Marketing targeting for outbound + paid

The 8 ICP dimensions

Firmographics — industry, vertical, size, geography, age
Tech-stack signals — what they use today (and what's missing)
Buyer persona — role, level, tenure, team size
Jobs-to-be-done (JTBD) — what they're trying to accomplish
Existing alternatives — how they solve the JTBD today
Trigger events — why now (new funding, new exec, regulation, etc.)
Budget authority — who controls spend
Reachability — can we find + contact them via channels we have

ICP vs Persona vs Segment

ICP = the company/account (firmographics + tech + situation)
Persona = the human buyer (role + motivations + objections)
Segment = a cluster of similar ICPs (often by size or vertical)

Use ICP for "which companies." Use persona for "which humans within ICP."

Clarify First

Before defining the ICP, confirm these inputs. If any is unknown or vague, ASK — do not assume:

[ ] Existing customers vs greenfield — refine from closed-won/best-customer patterns vs design-partner proxy (the real ICP comes from data, not a guess)
[ ] The JTBD the product solves — what customers are trying to accomplish (anchors firmographics, alternatives, and trigger events)
[ ] Reachable channels you have — paid / outbound / events / content / partners (an ICP you can't reach is academic; drives the Reachability dimension)

Stop rule: ask only the 2-3 that most change the output. If the user says "just draft it," proceed and list your assumptions at the top of the artifact.

Workflow

Step 1 — Start from your best customers

If you have customers:

List top 20 by health, revenue, expansion
Find common patterns across them
That's your real ICP (not what you guessed)

If you don't have customers:

Use design partners / pilot customers as proxy
Be ready to revise after 10-20 real customers

Step 2 — Disqualify common myth-ICPs

Common "ICP" patterns that are wishful thinking:

"Companies with 100-10,000 employees" — too broad
"Innovative companies" — meaningless
"Companies that want to grow" — everyone
"Enterprise" — undefined; map to size

A real ICP excludes most companies.

Step 3 — Document the 8 dimensions

Per dimension, be specific.

Example for HR analytics SaaS:

| Dimension | Definition | |-----------|------------| | Firmographics | US, mid-market (200-2000 EE), SaaS or services vertical, > $20M revenue, 5+ years old | | Tech-stack | Workday or BambooHR + ADP or Gusto. Bonus: existing BI tool (Looker / Tableau / Mode) | | Buyer persona | HR Director or VP People; 5+ years tenure; built career on people analytics | | JTBD | "I need to be a strategic partner to the CFO/CEO; I'm stuck doing reports manually" | | Existing alternatives | Excel + analyst (often resigned to it); occasionally hired contractor | | Trigger events | Annual reporting cycle just done; new CFO joined; HR analyst departed; board asking for better data | | Budget authority | Director can recommend; VP People approves $50K; CHRO approves $200K+ | | Reachability | SHRM events, HR Tech podcast, LinkedIn (HR Director groups), HR Brew newsletter |

Step 4 — Build the qualification rubric

For sales qualification, distill ICP to a scorable checklist:

Qualification (BANT-style):
- Industry fit: SaaS/services? (yes/no)
- Size fit: 200-2000 EE? (yes/no)
- Tech fit: Workday/BambooHR + ADP/Gusto? (yes/no)
- Authority: VP People+ in deal? (yes/no)
- Pain: actively trying to solve HR analytics? (yes/no)
- Budget: confirmed > $50K? (yes/no)
- Timeline: decision in next 6 months? (yes/no)

5+ yes = strong lead. <3 yes = disqualify.

Step 5 — Run `icp_scorer.py`

Audit ICP definition for specificity, score account lists against ICP.

python3 project-management/gtm/ideal-customer-profile/scripts/icp_scorer.py \
  --input icp_spec.json --format markdown

Step 6 — Refresh per quarter

ICP shifts:

New customers reveal new patterns
Product evolution opens new segments
Market shifts (recession, regulation, competition)

Don't refresh weekly. Quarterly is healthy.

Decision frameworks

Firmographic specificity test

For each dimension, ask:

Could I send a list of companies matching this to my BDR tomorrow?
If not, sharpen.

Vague: "growing SaaS companies" Sharp: "US-headquartered B2B SaaS companies, $10M-$100M ARR, post-Series B, that have hired their first VP of Sales in last 12 months"

Tech-stack signals

Tech stack reveals readiness:

Have a CRM = likely structured sales process
Have an HRIS = HR team beyond founder/HR-of-1
Use a specific vendor = open to category
Built in-house = NIH (Not Invented Here) bias risk

Tools like BuiltWith, G2, Crunchbase + scraping can reveal stack.

Trigger event types

Strongest triggers (cause buying activity):

New executive in role (60-day window)
Funding round closed (have budget; want to spend strategically)
M&A activity (integration needs)
Regulatory deadline (compliance pressure)
Major incident (urgent need)
Vendor consolidation push
Annual planning cycle (Q4)
Reorg / department restructure

Weaker triggers:

"They might want it"
"They have the budget"

Reachability test

For each ICP, can you:

Find them via paid acquisition? (size of audience; CPC)
Find them via outbound? (LinkedIn search, ZoomInfo, Apollo)
Find them at events? (SHRM, RSA, AWS reInvent, etc.)
Find them via content? (SEO keywords they search)
Find them via partners? (which adjacent vendors)

If unreachable, ICP is academic.

Common engagements

"Define ICP for our new product"

Pull design partner / pilot data (or interview 5-10 ideal targets).
Document the 8 dimensions.
Build qualification rubric.
Validate with sales: can they recognize these in inbound?
Validate with marketing: can they target this list?

"Refine ICP from closed customers"

List top 20 paying customers by health/expansion.
Find patterns: size, vertical, role, trigger, tech stack.
Document refined ICP.
Compare to original ICP: what's different?
Update qualification rubric + targeting.
Sunset segments not in refined ICP.

"Pipeline conversion is bad — is it ICP?"

Audit closed-won vs closed-lost ICP fit scores.
Common failure: pipeline is full of "near-ICP" that don't convert.
Tighten qualification; reject more aggressively at intake.

Anti-patterns to avoid

ICP = "everyone." Diluted GTM.
ICP without trigger event. Targets without urgency.
ICP unreachable via your channels. Academic exercise.
ICP that doesn't disqualify anyone. Not really an ICP.
Same ICP for 3 years post-launch. Customers reveal real ICP; refresh.
ICP defined by product, not customer. "Companies that need X" = solution-thinking.
ICP without qualification rubric. Sales can't apply it.

References

references/icp-dimensions-deep.md — 8 dimensions in depth + signals
references/icp-refinement-from-data.md — using closed customers to refine

Related skills

project-management/gtm/gtm-strategy — uses ICP as input
project-management/strategy-frameworks/business-model-canvas — segments block
project-management/discovery/customer-interview-script — interview-based ICP discovery
marketing/competitive-teardown — competitive context
c-level-advisor/cro-advisor — sales context
c-level-advisor/cmo-advisor — marketing context

Agent Skills: Ideal Customer Profile (ICP)

Install this agent skill to your local

Skill Files