LeadScorerWhat is an ideal customer profile?
An ideal customer profile, or ICP, describes the type of organization that is a strong match for a specific product or service. It identifies companies that can benefit from the offer and are commercially and operationally sensible to serve. In B2B, an ICP describes the company; buyer personas describe the people involved in the decision. [1] [2] [3]
This template turns that definition into a working decision system: which accounts to research, what evidence to look for, whom to approach, and what to rule out. The framework and example score weights below are LeadScorer’s proposed working model—not a published industry standard.
“Growing companies that need more revenue.”
“B2B marketing agencies with 10–50 employees, a focused service offer, a sales owner, and capacity to serve new clients.”
Illustrative boundaries; verify with customer evidence.ICP, buyer persona, intent, and qualification are different decisions
| Concept | Question it answers | What it does not prove |
|---|---|---|
| Ideal customer profile | Which organizations could be a good fit? | That a particular company is buying now. |
| Buyer role / persona | Who experiences, evaluates, influences, or approves the purchase? | That a job title establishes actual authority. |
| Trigger / engagement signal | What changed, or what relevant behavior occurred? | That there is an active, funded project. |
| Target account list | Which named companies should we research or work? | That every listed company has passed qualification. |
| Qualified conversation / opportunity | What has the buyer actually confirmed, and what next step is justified? | That fit or interest alone guarantees a sale. |
The company/person distinction is supported by HubSpot. HubSpot and Oracle also distinguish fit from engagement; MEDDICC distinguishes a buyer’s authority from a contact’s title. The table’s operational boundaries are this template’s synthesis. [1] [4] [5] [6]
The B2B ICP worksheet
Use one version per materially different offer or buying situation. Each important rule should include a source, a checked date, and an honest label: hypothesis, observed fact, buyer-confirmed, or outcome-supported. A source reference is an audit trail—not independent verification.
01 / Offer and customer success
We sell: [specific offer]. The customer outcome is: [measurable change]. The current alternative is: [incumbent, internal work, or doing nothing]. The commercial model is: [price, currency, term, onboarding]. A worthwhile customer is one we can: [serve sustainably].
02 / Target account
Segment: [2–4 words]. Industry / subsector: [definition]. Business model: [type]. Buying unit: [parent, subsidiary, site, or team]. Serviceable geography: [location and what it refers to]. Scale: [employee, revenue, workload, or site range—only where useful]. Why scale matters: [reason].
03 / Problem and delivery fit
Problem / job: [specific workflow]. Observable evidence: [facts a researcher can find]. Cost of inaction: [consequence to verify]. Prerequisites: [systems, skills, data, materials, capacity, or certifications]. Proof needed: [evidence the buyer will evaluate].
04 / Buying roles
Entry contact: [function and title variants]. Problem owner: [role]. Potential champion: [role; behavior to verify]. Budget approver: [role; authority to confirm]. Evaluators: [technical, operational, finance, procurement, or legal]. Decision criteria: [what each role needs to believe].
05 / Signals and timing
Relevant event: [event or none observed]. Source and observed date: [reference]. Why it might matter: [hypothesis]. Review date: [context-specific]. Question that verifies timing: [direct question].
06 / Boundaries and qualification
Must-haves: [true eligibility gates]. Negative ICP: [permanent mismatch]. Not now: [temporary issue and review condition]. Suppress from outreach: [ownership, conflicts, opt-outs, or channel restrictions]. Before a handoff: [company, person, problem, interest in a next step, and commercial fit].
07 / Evidence and validation
Supporting evidence: [customer / CRM / research references]. Contradictory evidence: [losses, churn, delivery problems, exceptions]. Unknowns: [assumptions]. Test: [cohort, comparison, time window, decision rule]. Outcomes: [sales acceptance, opportunities, wins, economics, retention]. Owner / version / review: [details].
How to create an ICP that your team can actually use
Start with a specific offer—not a list of attractive companies
State the problem you solve, the result you deliver, the alternative you replace, and the circumstances in which your approach is useful. A software subscription, a machining contract, and a managed service need different delivery and commercial conditions. “Companies with budget” is not enough.
Study good outcomes and bad outcomes
Review customers that bought and succeeded, plus lost deals, no-decisions, churn, and difficult delivery. HubSpot’s research guidance includes CRM analysis, interviews, sales and support feedback, and unhappy customers—not only your advocates. [9]
In this worksheet, “best customer” means more than the largest contract. Compare realized value, time to value, retention or repeat orders, expansion, delivery effort, and contribution after direct service costs. Cognism’s ICP guidance emphasizes value for both parties; Salesforce also includes long-term value and refinement. [2] [3]
Avoid hindsight leakage. Use retention and margin to judge whether a segment worked. Do not pretend those future outcomes were facts available before the prospect bought. Turn them into pre-purchase hypotheses—such as workflow compatibility or available implementation capacity—and test those hypotheses separately.
Interview for the actual buying situation
Use the following original question set with customers, lost prospects, and delivery colleagues. Ask for specific events and examples rather than agreement with your pitch.
What was happening when this became a problem? What had you already tried? What did leaving it unresolved cost you? Why did the issue become a priority then? Who first raised it, who evaluated alternatives, and who approved the purchase? What nearly stopped the decision? What happened after implementation? What would have made you a poor fit?
Separate what the interviewee observed from what they inferred. Preserve short, permissioned buyer language in your evidence log and record the date and context. A handful of interviews can improve a hypothesis, but it is not statistical proof of a market-wide rule.
Translate each important belief into a researchable criterion
For every criterion, specify the value, why it matters, where to find it, and what to do when it is unknown. “Has a complex finance operation” might become “several operating entities with a shared AP function,” with workflow complexity still checked in discovery. Revenue estimates and headcount bands are useful only when they are defensible proxies—not substitutes for need or buying capacity.
| Criterion | Research evidence | Still needs confirmation |
|---|---|---|
| Relevant company activity | Product, service, or business-unit pages | The exact use case and purchasing entity |
| Scale / operating complexity | Disclosed size, sites, team, or workload | Actual workload and delivery implications |
| Compatible environment | Documentation or dated system references | Current configuration and mandatory requirements |
| Recent change | A dated announcement or buyer statement | Whether the change creates a relevant project |
| Commercial fit | Your price and delivery model; buyer-provided context | Budget authority, buying process, and workable economics |
These are research suggestions, not guarantees that a particular data vendor has a field or that a public page proves current need.
Map roles without inventing authority or advocacy
The initial contact, day-to-day problem owner, evaluator, and final budget approver may be different people. MEDDICC defines the economic buyer through buying authority and a champion through influence and willingness to advocate—not friendliness alone. Record potential roles first, then confirm them through the actual purchase. [6] [7]
6sense’s 2025 buyer research also examines buying groups and evaluation behavior. Its sample involved purchases of at least $25,000, with a median between $200,000 and $300,000. That supports attention to buying context; it does not justify applying its group sizes or cycle lengths to every small agency retainer. [8]
Test the profile before expanding the campaign
Start with a clearly defined pilot you can research well. Compare the proposed segment with a sensible adjacent segment, keeping the offer, execution, and observation period as comparable as possible. Record attempted accounts and exclusions. Follow the same account cohort from response through sales acceptance, opportunities, wins, and—when enough time has passed—customer outcomes.
Do not change the ICP because one message performed poorly. Delivery problems, weak offers, uneven follow-up, and immature sales cycles are alternative explanations. Log raw counts and time windows; small samples are directional. Agree in advance what evidence would narrow, broaden, or retire a criterion. This is the validation approach recommended for this template, not a universal experiment design.
Ideal customer profile examples for B2B
Every example below is illustrative. The seller is shown first and the customer organization second. None is a verified LeadScorer client profile, a benchmark, or a promise of results.
Managed outbound → B2B marketing agencies
Test agencies with a focused offer, credible proof, a sales owner, and capacity to serve new clients. A sales hire is a research cue—not proof they need outsourcing. Exclude B2C-only needs and expectations of guaranteed closed revenue.
AP software → multi-entity service businesses
Look for a relevant invoice-approval workflow, supported systems, and an implementation owner. Map the controller, AP lead, IT evaluator, and actual budget approver. Exclude mandatory integrations the product cannot support.
Precision machining → industrial OEMs
Research the parts and sourcing situation. Materials, tolerances, certification, qualification, capacity, and viable production runs matter more than a generic employee range. Unsupported specifications are a gate, not a small score deduction.
Management reporting → B2B service businesses
Target a recurring reporting need with accessible financial data and a clear internal owner. A growing team may be a clue; the reporting gap must be confirmed. Exclude work beyond the firm’s professional and delivery capability.
Customs coordination → repeat industrial importers
Define cargo, port coverage, shipment pattern, operating capability, and service scope. A new sourcing lane may justify research. It does not establish dissatisfaction with the current broker or permission to contact anyone.
How to score ICP fit without confusing it with buying intent
HubSpot supports distinct fit and engagement scores; Oracle describes a two-dimensional approach that links profile fit and interest to different follow-up actions. This resource uses the same separation, but the weights and thresholds below are editable starter assumptions. They have not been statistically validated. [4] [5]
First, check suppression and hard eligibility. Then score five fit dimensions. A confirmed mismatch with a mandatory requirement cannot be rescued by a high total.
| Fit dimension | Starter weight | What to evaluate |
|---|---|---|
| Problem / use-case fit | 30 | The offer can address the relevant job or workflow. |
| Industry / business-model fit | 20 | The account operates in the situation the profile defines. |
| Scale / complexity fit | 15 | Workload and operating structure suit the offer. |
| Delivery / technical fit | 15 | Implementation, systems, capacity, or specifications are supportable. |
| Economic fit | 20 | There is a credible path to mutual economic value. |
For each dimension, use 0 = mismatch, 1 = partial fit, 2 = fit. Leave unknowns blank. Attach evidence references. Assess industry and scale independently only when they add distinct information; avoid rewarding the same underlying characteristic twice.
Assessment coverage = weight of rated dimensions ÷ total weight
Possible ceiling = entered points + weight of unknown dimensions
An unknown is not a confirmed mismatch, but it does not earn points. The workbook shows both coverage and the possible ceiling instead of inflating a score by ignoring unknown dimensions. A complete score is still a judgment based on entered evidence—not a probability of purchase.
The starter routing is: incomplete research → investigate; complete fit of 80–100 → priority-fit segment; 60–79 → adjacent test segment; below 60 → lower priority. Suppression, failed gates, missing evidence references, invalid inputs, and temporary holds override this routing. Actual timing is recorded separately; even a high-fit account with an active project still needs discovery and qualification.
The mistakes this template is designed to prevent
Invented precision. A narrow revenue range is not useful merely because it looks specific. Record why the range matters and how it was established.
Revenue-only selection. A large contract with poor delivery economics is not automatically the model to replicate. Check customer outcomes and the work required to produce them.
Survivorship bias. A profile based only on visible success stories ignores churn, no-decisions, difficult customers, and potential customers your current channel never reached.
Confusing a signal with a diagnosis. Hiring, funding, or opening a site may justify a question. It does not prove a problem, intent, budget, or openness to a provider.
Turning temporary timing into a permanent exclusion. “Not this quarter” and “we cannot meet this specification” require different actions. “Do not contact” is different again.
Calling a finished form validated. Form completion means the draft has been documented. Validation requires comparing its predictions with real account and customer outcomes.
Ideal customer profile template FAQs
What should a B2B ICP template include?
Include the offer, company characteristics, use case, delivery and economic fit, buying roles, timing signals, exclusions, evidence, and a validation plan. The goal is an actionable account definition—not a fictional biography.
How is an ICP different from a buyer persona?
The ICP describes the organization. A buyer persona describes a type of person involved in the buying decision, including their responsibilities and motivations. Use both, but do not replace company fit with personal demographics. [1]
Can I create an ICP without existing customers?
Yes. Begin with an explicitly labeled hypothesis based on the problem you can solve, your delivery constraints, interviews, and observable market evidence. Use a narrow pilot to learn. Do not describe a research-backed hypothesis as a validated customer pattern.
How many ICPs should I create?
For this template, start with one per materially different offer or buying situation. Split profiles when qualification, economics, delivery, or the buying process genuinely changes—not merely to give every industry a separate document.
Does a high ICP score mean a lead is sales-qualified?
No. It means the entered evidence supports fit under your chosen rules. Confirm the relevant person, problem, interest in a next step, and commercial situation. Keep fit, timing, and opportunity qualification separate. [5]
Should revenue and headcount always be required?
No. Use them when they help explain workload, complexity, economics, or delivery fit. A disclosed company size is not proof of a budget. Missing estimates should trigger research, not invented numbers.
Can AI generate the ICP for me?
AI can help organize supplied evidence and draft hypotheses. This builder is a transparent form-based tool: it does not research companies, enrich contacts, or invent missing facts. Review every important rule against real evidence before relying on it.
How often should I update the profile?
Assign an owner and a review date. Revisit it after meaningful changes in the offer, pricing, delivery capability, market, or customer outcomes. Review cadence depends on your learning cycle; calendar updates alone do not make a profile more accurate. [2]
Is this tool free, and where does my data go?
The supplied builder has no signup, API calls, or tracking. Your entries stay in the current browser session unless you choose to save locally or export. A local save is not a secure vault: use anonymized business references, not confidential personal or customer data.
What can I download or export?
The download is an editable .xlsx workbook with formulas and instructions. The builder also exports a Markdown brief, a plain CSV field list, a portable JSON draft, and a printer-friendly version you can save as a PDF using your browser. The workbook and webpage are separate tools; changes in one do not automatically update the other.
Research sources and scope notes
These sources inform the distinctions and working method in this resource. The template’s custom prompts, score weights and routing thresholds remain LeadScorer’s editable working model—not external benchmarks or promises of performance.
Ideal customer profiles and buyer personas: How are they different? ↗
Used for: Separate the organization-level ICP from the people involved in buying.
Limit: Practitioner guidance, not a causal experiment. This template deliberately excludes irrelevant personal demographics.
Ideal Customer Profiles (ICPs): Benefits & How to Create ↗
Used for: Use company attributes, customer outcomes, customer feedback, and ongoing refinement.
Limit: General vendor guidance. No claimed conversion or retention uplift is adopted as a benchmark.
How to Create an Ideal Customer Profile (ICP) With Template ↗
Used for: Look for mutual value: the buyer benefits and the seller can serve the account economically.
Limit: Commercial vendor article. Its secondary marketing statistics are not used.
Overview of the lead scoring tool ↗
Used for: Keep fit, engagement, exclusions, and configurable scoring rules distinct.
Limit: Product documentation supports a design pattern; it does not validate the custom weights in this resource.
What is Lead Scoring? ↗
Used for: Separate identity/fit from interest, map scores to actions, and recalibrate using sales feedback.
Limit: Use conceptual guidance only. Illustrative thresholds in this resource are not Oracle benchmarks.
Economic Buyer ↗
Used for: Map the role with overall buying authority rather than assuming the highest title is the buyer.
Limit: A role hypothesis is not verified authority. Confirm authority in the specific purchase.
Champion ↗
Used for: Distinguish a helpful contact from someone who can and will advocate internally.
Limit: An initial contact should be labeled a potential champion until behavior supports that role.
2025 B2B Buyer Experience Report ↗
Used for: Account for buying groups, prior alternatives, evaluation, and implementation concerns.
Limit: Vendor-sponsored, self-reported research. Main survey: nearly 4,000 respondents; minimum purchase $25,000; median $200,000–$300,000. Do not generalize its cycle lengths or group sizes to every small B2B service deal.
How to create detailed buyer personas for your business ↗
Used for: Combine interviews, CRM patterns, sales/support feedback, and unhappy-customer evidence.
Limit: Adapt the research method to B2B organizational decisions, not invented ages, hobbies, or personality traits.
Understand JavaScript SEO basics ↗
Used for: Keep useful editorial content, crawlable links, title, and canonical available independently of form interaction.
Limit: Technical implementation guidance, not a ranking guarantee.
B2B Outbound Resources & Free Tools ↗
Used for: Preserve the ungated resource promise and connect the builder to existing outbound resources.
Limit: Website context only; not independent evidence of effectiveness.