ICP Builder
Input your best customers and generate your Ideal Customer Profile in minutes. AI analyzes patterns across your top accounts to define who you should target.
Your Best Customers
Add at least 3 of your best customers. The more you add, the more accurate your ICP.
How to build an Ideal Customer Profile that actually works
An Ideal Customer Profile is the most leveraged document a revenue team owns and the most commonly ignored. Done well, it disqualifies bad-fit prospects before sales spends an hour on them, focuses marketing spend on the segments most likely to convert, and gives product the customer pattern that should drive the next quarter's roadmap. Done badly — or written once and never revisited — it becomes a deck nobody reads.
The framework below is what the tool above is built on. The five fields it generates — industry, company size, buyer role, pain points, deal size — cover the targeting basics. What turns those fields into something useful is how you choose your input customers and how you wire the resulting profile into your qualification rules.
1. Pick the right input customers
The customers you analyze determine everything. Use customers who renewed at full price, expanded their contract, or referred new business. Skip churned accounts, deeply discounted deals, and one-off projects. The signal you want is “repeatedly chose us and got value” — not “signed once and disappeared.”
Three is the floor for finding patterns. Five to ten is the sweet spot. Past ten you're usually averaging across segments that should be split into separate ICPs anyway. If your top customers fall into clearly distinct groups (e.g., agencies vs. in-house teams, SMB vs. mid-market), build separate profiles per segment rather than averaging them.
A common mistake: using your loudest customers. The customer who emails you the most feedback isn't necessarily your best fit — they're sometimes your worst-fit customer fighting against the product. Lean on quiet, expanding accounts. They tell you what the product is actually solving for the people you want more of.
2. ICP vs. buyer persona — they're different
An ICP describes the company. A buyer persona describes the person inside that company. They're complementary and you need both. The ICP guides targeting, qualification, and account selection. The persona guides messaging, channel choice, and content.
The tool above outputs ICP fields plus a buyer-role list. Most B2B sales motions touch 2–4 personas inside a single ICP company: the user (who feels the pain), the champion (who advocates internally), the economic buyer (who signs), and the gatekeeper (security, IT, procurement). Map each one once you have the ICP locked.
3. Make the ICP disqualifying, not just descriptive
Most ICPs read like marketing copy: “mid-market SaaS companies looking to scale.” That tells sales nothing actionable. A useful ICP is specific enough to disqualify: “US-based B2B SaaS companies, 50–500 employees, post-Series-A, sales-led motion, average ACV $15K–$60K, with an existing CRM and at least one full-time SDR.” That version filters inbound automatically.
For each ICP field, ask: would I disqualify a lead that doesn't meet this? If yes, keep it. If no, the field is too vague to be useful — tighten it or remove it. The goal is a profile sales actually applies in the moment, not a poster on the wall.
4. Translate ICP into qualification rules
The bridge between an ICP document and pipeline impact is qualification questions. Each ICP field becomes one or two questions you ask every inbound lead:
- Industry → “What does your company do?”
- Company size → “How many people on your team?”
- Buyer role → “What's your role in this evaluation?”
- Pain → “What prompted you to look right now?”
- Deal size → discovered through scope and use-case questions
- Buying triggers → “Has anything changed recently?”
Asked as a form, these questions abandon at 50%+. Asked as a natural conversation across two or three back-and-forth replies, they complete at 80%+. That's the design principle behind conversational AI qualification: same data, less friction. Leadstr runs that conversation against your ICP automatically — every inbound lead, 60-second first response, full data written to your CRM.
5. Update your ICP on a real cadence
ICPs go stale. Your product changes, your pricing changes, your competition changes, and the customer set that fit you 18 months ago isn't necessarily who you should target today. Quarterly during periods of fast growth, annually otherwise. Update sooner if you ship a major product change, enter a new segment, or notice your win rate or churn rate moving more than 10 points.
ICP — common questions
What is an Ideal Customer Profile (ICP)?
An ICP is a written description of the company and buyer most likely to get value from your product, evaluate quickly, and become a long-term customer. It typically includes industry, company size, buyer role, pain points, deal size, and buying triggers.
ICP vs buyer persona — what's the difference?
An ICP describes the company you want to sell to. A buyer persona describes the individual inside that company. You need both: ICP for targeting and qualification, persona for messaging.
How many customers should I input?
Three is the minimum for the tool to find patterns. Five to ten is the sweet spot. Beyond ten you'll see diminishing returns unless your customers span very different segments.
Should I include my best customers, my newest, or my biggest?
Use customers who have stayed past renewal or referred others — those signal real fit. Avoid one-time projects, churned accounts, and customers who got discounted into the deal. Quality of fit beats deal size.
How often should I update my ICP?
Quarterly during high growth, annually after that. Update sooner if you ship a major product change, enter a new segment, or notice your win rate shifting.
How does Leadstr use this ICP?
Once you have an ICP, Leadstr scores every inbound lead against it in real time — checking industry fit, role match, deal size band, and pain-point alignment during the qualification conversation.
When this builder won't produce useful output
- You don't have 3+ paying customers yet. Pattern-matching across <3 inputs produces noise, not signal. Pre-product-market-fit teams should interview prospects instead.
- Your customers are wildly different from each other. If your top 10 customers span SMB B2C to enterprise B2B, averaging them creates a fictional persona. Segment first, build separate ICPs.
- You sell to consumers (B2C). The fields here (company size, buyer role) don't translate cleanly to consumer purchase decisions. B2C personas need different inputs.
- You want to expand into a new segment you've never sold to. ICP analysis works on past data — it can't tell you about a market you haven't entered yet.