> ## Documentation Index
> Fetch the complete documentation index at: https://docs.bitscale.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# F-2: Understanding Your ICP

> How to define an ICP that's specific enough to build a targeting system around — not just a Confluence doc nobody reads.

<Info>
  **F-2 · GTM Foundation · 100 XP · \~18 min**
</Info>

Most ICP documents are wrong. Not because the team didn't think hard enough — but because they defined the ICP based on who they want to sell to, not who actually buys. This module teaches you the difference, and how to translate a real ICP into targeting criteria you can automate.

***

## Why Most ICPs Are Wrong

The classic ICP document lists things like: "B2B SaaS companies, 50–500 employees, Series A–C, US-based." That describes half the companies on LinkedIn. It's not a targeting definition — it's a demographic sketch.

A real ICP answers a different question: **what combination of attributes predicts that this company will buy from us, get value, and stay?**

The answer always comes from your existing customers. Not from a whiteboard.

***

## Building an ICP from Existing Customer Data

Run this analysis on your best customers — define "best" as highest LTV, fastest time-to-value, or fewest support tickets:

| Attribute               | Where to Find It         | What to Look For                        |
| ----------------------- | ------------------------ | --------------------------------------- |
| **Company size**        | LinkedIn, Clearbit       | Is there a range that clusters?         |
| **Industry / vertical** | CRM, LinkedIn            | Which verticals appear most?            |
| **Funding stage**       | Crunchbase, Apollo       | Bootstrap vs VC-backed pattern?         |
| **Tech stack**          | BuiltWith, Clearbit      | What tools do they all use?             |
| **Team structure**      | LinkedIn job postings    | SDR team? Marketing team size?          |
| **Growth signals**      | LinkedIn headcount trend | Were they hiring when they bought?      |
| **Trigger event**       | CRM notes, deal history  | What made them look for a solution now? |

When you map 20–30 of your best customers against these columns, patterns emerge. Those patterns are your real ICP.

***

## The ICP Scoring Matrix

Once you've identified patterns, translate them into a scoring rubric:

```
ICP Fit Score (0–10):

Firmographic fit:      0–3 points
  - Industry match:    +1
  - Size in range:     +1
  - Funding stage:     +1

Technographic fit:     0–3 points
  - Uses CRM:          +1
  - Has SEP (e.g. Outreach/Instantly): +1
  - Has data tool (Apollo/ZoomInfo):   +1

Signal fit:            0–4 points
  - Actively hiring SDRs: +2
  - Recent funding:       +1
  - Headcount growth >20%: +1

Score 8–10: Hot ICP — top priority
Score 5–7:  Warm ICP — work the signal
Score 0–4:  Outside ICP — deprioritize
```

This matrix becomes the backbone of your enrichment and scoring columns in Bitscale.

***

## Negative ICP: Who to Exclude

Equally important: who you should not spend time on.

Common negative ICP signals:

* Fewer than 10 employees (no budget, no infrastructure)
* No sales team at all (wrong use case)
* Competitor's customer (already solved the problem)
* Wrong geography (if you have regional constraints)
* Stage too early or too late (pre-product or enterprise with 18-month sales cycle)

Build your negative ICP list explicitly. In Bitscale, you can filter these out before they ever reach your outbound queue.

***

## Encoding ICP as Bitscale Criteria

Once you have your ICP defined, it lives in your Bitscale grids as column filters and scoring formulas:

* **Industry** column → filter to target verticals
* **Employee count** column → filter to size range
* **Tech stack** columns → flag CRM/SEP presence
* **Hiring signal** column → detect SDR hiring
* **ICP score** formula column → sum the above

You'll build this in F-3 (Your First Grid). For now, make sure your ICP is specific enough to encode — if you can't write it as a filter, it's not specific enough.

***

<Tip>
  **Quick Check:** Can you name three firmographic and two technographic criteria that predict a customer buys from you? If you can't yet, that's what the F-2 challenge will help you figure out.
</Tip>

***

## F-2 Challenge: Build Your ICP Scoring Matrix (+100 XP)

Take your 10 best current customers. Map them against the attributes above. Identify the top 3 patterns that appear across 7+ of them. Write up your ICP scoring matrix.

<Card title="Submit F-2 Challenge →" icon="upload" href="https://bitscale.fillout.com/academy-challenge-f2">
  Submit your ICP scoring matrix based on real customer data. **+100 XP on approval.**
</Card>

***

<Card title="Next: F-3 Your First Grid →" icon="arrow-right" href="/academy/foundation/your-first-grid">
  Turn your ICP definition into a live Bitscale grid that finds and scores prospects automatically.
</Card>