ICP and Buying Signals: The Complete B2B Targeting Guide

TL;DR

An ideal customer profile (ICP) defines who to target: the companies most likely to buy, get value from your product, and stick around. Buying signals define when to act — the funding events, hiring patterns, and tech changes that show a company is actually in-market right now. Most teams build one of the two and stop: a static ICP list with no sense of timing, or a stream of buying signals with no filter for whether the account was ever a fit in the first place.

Neither works well alone. This guide covers how to build a real ICP from your existing customer data, what "persona-level insight" actually adds on top of it, how to combine fit and signal into one targeting framework, and where that framework should change what you actually send.

What Is an Ideal Customer Profile?

An ICP is a structured, company-level description of the accounts most likely to buy, expand, and retain. It's built from evidence, not from a wishlist of dream logos. It's easy to confuse with two adjacent concepts that answer different questions:

Concept	Question It Answers	Level
Target market	How big is the addressable opportunity?	Market
Ideal customer profile	Which companies are worth pursuing?	Company
Buyer persona	Which individuals inside that company matter, and why?	Individual

Target market is the widest lens, ICP narrows it to specific companies, and persona narrows it further to the people inside those companies you actually need to convince. A common mistake is building a persona and calling it an ICP, or building an ICP and assuming it tells you anything about the individual buyer — it doesn't.

What Are Persona-Level Insights Within an ICP?

Persona-level insight is the individual layer sitting inside your company-level ICP: title, seniority, role in the buying committee, and the specific pain point that person cares about. A company can be a perfect ICP fit on paper: right industry, right size, right tech stack. Outreach still fails if it's aimed at the wrong person inside that account, or the right person with a message that ignores what they specifically care about.

This is where an ICP stops being a filtering tool and starts being a targeting tool. The account-level criteria tell you which companies belong on the list. The persona-level detail tells you who to actually contact within each one, and what angle to use once you've reached them. A VP of Sales and a Head of RevOps at the same ICP-fit company are worth reaching with different messages, even though the company itself scores identically on fit.

How Do You Build an ICP From Scratch?

1. Pull your closed-won data — export your best 20-50 customers, defined by revenue, retention, and expansion, not just recognizable logos
2. Find the shared attributes — look for the patterns that repeat: industry cluster, headcount band, common tech stack, recurring trigger event that preceded the deal
3. Build the anti-ICP — list the customers who churned, haggled endlessly, or drained support disproportionately; the traits they share are your disqualifiers
4. Codify it into scoreable criteria — turn the patterns into concrete, filterable rules rather than a paragraph in a slide deck
5. Revisit it quarterly — a profile built once and never updated drifts out of date as your product and market shift

The profile is only useful once every team can apply it the same way. If sales, marketing, and product each have a slightly different idea of what counts as ICP-fit, none of them are actually using it as a filter.

What Data Layers Make Up a Complete ICP?

Layer	What It Captures	Example
Firmographic	Industry, company size, geography	Mid-market SaaS, 50-500 employees, North America
Technographic	Tech stack and tools already in use	Uses Salesforce, HubSpot, or a specific competitor
Behavioral / trigger	Hiring, funding, growth events	Recently raised a Series B, hiring 3 SDRs
Organizational readiness	Buyer persona and process maturity	Has a dedicated RevOps function
Negative indicators	Traits that disqualify despite surface fit	Long procurement cycles, history of high churn in this segment

Firmographics alone describe a company. The other four layers are what actually predict whether it will buy, expand, or become the kind of account that drains support without ever renewing.

What's the Difference Between ICP Fit and Buying Signals?

Fit and signal answer two different questions, and conflating them is where most targeting frameworks break down.

Fit is who. It's built from stable, slow-changing attributes: industry, headcount, tech stack, geography. An account's fit score doesn't move much week to week.

Signal is when. It's built from fast-changing, time-sensitive events: a funding round, a new VP hire, a tech stack change, active research behavior. The same account can go from no signal to a strong signal overnight.

Treating these as one thing causes two distinct failure modes. Chasing every signal without checking fit first fills the pipeline with in-market accounts that were never going to be a good customer. Sitting on a well-built ICP list with no signal layered on top means reaching accounts at an arbitrary moment instead of the moment they're actually ready. Timing is a big part of why signal-based personalization outperforms generic, scheduled outreach.

How Do You Combine ICP Fit With Buying Signals for Targeting?

Score the two dimensions separately, then route accounts based on where they land:

	Strong Buying Signal	Weak or No Signal
High ICP Fit	Engage now — highest priority	Add to a watch list, monitor for new signals
Low ICP Fit	Verify before committing rep time	Deprioritize

Fit determines whether an account belongs on your target list in the first place. Signal determines when to actually spend rep time on it. An account with strong signals but poor fit isn't a shortcut worth taking. It's usually a sign to verify the signal before treating it as sales-ready, not a reason to skip the fit check.

Common Mistakes When Combining ICP and Signals

• Acting on a single signal in isolation — one data point (a job change, a funding mention) can mislead on its own; corroborating signals are far more reliable than any single trigger
• Skipping the fit check on inbound signal alerts — a strong buying signal from an account outside your ICP shouldn't get the same urgency as the same signal from a genuine fit
• Building the ICP once and never revisiting it — your first version of the profile will be wrong in at least one dimension, and product or market shifts make an unreviewed ICP stale within a couple of quarters
• Ignoring the anti-ICP entirely — disqualifying traits matter as much as qualifying ones; a company that matches every positive criterion but also matches a known churn pattern isn't a clean win

How Does This Change What You Actually Send?

Here's the part that determines whether all of this work shows up in a reply rate or just in a CRM field: the same buying signal should produce a different email depending on which ICP segment it's attached to. A funding announcement means something different to a company you'd classify as an early-adopter fit than to one in your core enterprise segment. The outreach should read that way, not use the same template with the company name swapped in.

Signal-based cold email personalization works at exactly that intersection: reading the fit context and the specific signal together, then writing outreach that reflects both, instead of a generic template applied uniformly across an otherwise well-built target list. A strong ICP and a reliable signal feed get you to the right account at the right moment. What you actually say once you're there is a different problem, and it's the one Flinter handles: turning a fit-plus-signal match into copy that reads like someone did the homework, not a mail merge with the right name in it.

---

Want to see how your own ICP and signal data translate into personalized outreach? Book a 30-minute walkthrough.