# ICP template for SaaS

> A fillable ICP with 22 fields across firmographic, technographic, behavioural and trigger criteria, plus a negative ICP and two completed examples at different ACVs.

Source: https://saas-marketing.net/templates/ideal-customer-profile/
Topic: SaaS Marketing
Type: template
Published: 2026-09-11
Last updated: 2026-09-11
Publisher: SaaS Marketing (saas-marketing.net)
License: CC BY 4.0. Quote or republish with attribution and a link to https://saas-marketing.net/templates/ideal-customer-profile/

## Short answer

A working SaaS ICP has 22 fields across four groups: firmographic (7), technographic (4), behavioural (6) and trigger (5). Each field carries a data source and a weight, and the weights produce an account fit score from 0 to 100 that lead routing and ad audiences read directly. Add a negative ICP listing the account types you refuse. Refresh it quarterly from closed won and churn analysis, not from opinion.

## Key takeaways

- An ICP that is not wired into lead routing and ad targeting is a poster, and posters do not change pipeline.
- The negative ICP section prevents more wasted spend than the positive one, because it kills demand you are already paying for.
- Score accounts 0 to 100 with weights, then set tier thresholds at 80 and 55 rather than treating fit as a yes or no.
- Trigger fields decay fastest and matter most, so they get refreshed monthly while firmographics can wait a quarter.
- Build the first version from your 20 best closed won accounts and your last 15 churned accounts, never from a persona workshop.
- Self serve and sales led products need different ICPs even inside one company, because the fit criteria barely overlap.

---

Most ICP documents are a slide with four bullets: mid-market, North America, ops teams, 50 to 500 employees. That definition matches roughly 400,000 companies, which makes it useless for routing a lead and worse than useless for building an ad audience.

The version below is a working document with 22 fields, a source for each one, and a scoring model that outputs a number your CRM can read. It ends with the operational steps, because an ICP that never reaches lead routing or a paid audience does nothing except make people feel organised.

## The 22 fields, grouped and sourced

Each field has an instruction line and a data source. Fields marked with a weight feed the scoring model in the next section. If you cannot populate a field from an existing source within a week, either find a proxy or delete the field.

**Firmographic (7 fields)**

| Field | Instruction | Source | Weight |
|---|---|---|---|
| Employee band | State a floor and a ceiling, not a midpoint | Closed won analysis | 10 |
| Revenue band | Use it only if your pricing scales with revenue | Enrichment | 4 |
| Industry or vertical | Name the two or three that win fastest, not ten that are possible | Closed won win rate by SIC | 10 |
| Geography | Include the markets you can legally and practically support | Contract and support data | 5 |
| Funding or ownership | Venture backed, PE owned, bootstrapped, public | Crunchbase or enrichment | 3 |
| Owning team exists | Is there a named team that owns this workflow today? | Sales call notes | 8 |
| Procurement threshold | Below what deal size does their process stay simple? | Won deal cycle time analysis | 4 |

**Technographic (4 fields)**

| Field | Instruction | Source | Weight |
|---|---|---|---|
| System of record | The tool you must integrate with to be usable | Integration usage data | 8 |
| Adjacent stack | Two or three tools whose presence predicts fit | Enrichment plus win analysis | 5 |
| Integration requirement | Which integration is a hard gate in evaluation? | Lost deal reasons | 4 |
| Data maturity | Do they already have the data your product needs? | Onboarding time by account | 3 |

**Behavioural and operational (6 fields)**

| Field | Instruction | Source | Weight |
|---|---|---|---|
| Volume metric | The unit that makes your product worth paying for, with a floor | Product usage by account | 12 |
| People in the workflow | Minimum number of humans touching the process | Seat count at renewal | 6 |
| Workflow frequency | Daily, weekly or monthly, and why that matters to retention | Usage frequency by cohort | 5 |
| Current solution | Spreadsheet, in house build, or a named competitor | Discovery call tagging | 4 |
| Budget owner seniority | Which title signs, and at what amount | Won deal signatory analysis | 3 |
| Security review depth | SOC 2 only, or full vendor assessment with a questionnaire | Deal cycle audit | 3 |

**Trigger (5 fields)**

| Field | Instruction | Source | Weight |
|---|---|---|---|
| New hire in owning role | The title whose arrival starts an evaluation | Job posting monitoring | 2 |
| Funding event | Round size and recency that correlates with buying | Funding alerts | 1 |
| Compliance or regulatory deadline | Named regulation with a date | Industry calendar | 2 |
| Incumbent renewal window | When their current contract expires | Discovery questions | 2 |
| Headcount growth rate | Growth threshold that breaks their current process | Enrichment deltas | 2 |

Weights total 100. The volume metric carries the heaviest weight on purpose, because in every SaaS ICP audit we have run it predicts both win rate and retention better than employee count does. For the conceptual background, see [ideal customer profile (ICP)](/glossary/ideal-customer-profile/), and for the market sizing that sits upstream of this document, read [how to define your SaaS target market](/guides/saas-target-market/).

## The scoring model and the three routing tiers

Score each field 0, 3 or 5, multiply by its weight, then divide by 5 to land on a 0 to 100 scale. Three tiers come out of it, and each tier gets a different treatment.

Check the thresholds rather than trusting them. If tier A accounts win at 31% and tier B at 27%, your weights are not discriminating and the model needs work. A healthy split usually shows tier A winning at two to three times the tier B rate. The [marketing qualified lead](/glossary/marketing-qualified-lead/) definition should reference this score directly, otherwise you end up with two competing definitions of a good lead and a permanent argument with sales.

## The negative ICP, which saves more money than the positive one

List the account types you refuse, with a reason and an enforcement mechanism. This section is short and it will be the most argued over.

**Typical negative ICP entries**

Enforcement matters more than the list. Every entry needs a mechanism: an exclusion list in the ad platform, a disqualification reason in the CRM, a form field that routes the account away. Without that, the negative ICP is advisory and gets ignored the first time a rep spots a logo they want.

## Two completed examples at different ACVs

The same template produces very different documents depending on price point. Both examples below are composites built from real accounts.

| Field | Self serve, $6K ACV | Mid-market, $75K ACV |
|---|---|---|
| Employee band | 10 to 120 | 400 to 3,000 |
| Industry | Digital agencies and studios | Financial services and healthcare |
| Volume metric | 25+ active client projects | 8,000+ monthly transactions |
| System of record | Slack and Notion | Salesforce and Workday |
| Owning team exists | Not required, one ops generalist is enough | Required, a named compliance or ops function |
| Budget owner | Founder or head of ops, card payment | VP or director, procurement involved above $50K |
| Security review | None expected | Full vendor assessment, SOC 2 Type II plus questionnaire |
| Primary trigger | Headcount passing 20 | Regulatory deadline or incumbent renewal |
| Sales motion | No sales touch below 15 seats | AE led, 90 to 120 day cycle |

Notice how little the two share. If your company runs both motions, write two ICP documents and two scoring models. Merging them produces a blended definition that describes neither, which is the most common failure we find when running the [SaaS marketing audit checklist](/checklists/saas-marketing-audit/) at companies with a hybrid motion.

Pull your 20 highest value closed won accounts from the last 18 months and your 15 most recent churned accounts. Find the fields where the two groups differ most. That difference is your ICP. A workshop full of opinions produces a document that flatters the product rather than describing the market.

## Where each field's data actually comes from

Four sources cover almost every field. Closed won analysis, exported from your CRM with product usage joined on, gives you firmographics, volume metrics and win rates by band. Churn analysis gives you retention predictors, which the acquisition focused version of an ICP always misses.

Call recordings give you the behavioural fields. Search Gong or your recording tool for the phrase your best customers use to describe the problem, then count how often it appears in won versus lost deals. Enrichment tools such as Clearbit, Apollo or Clay fill technographic and firmographic gaps at scale, though their accuracy on employee count in the 50 to 200 band is mediocre and worth spot checking against LinkedIn.

The trigger fields need monitoring rather than analysis. Job posting alerts for the owning role, funding alerts, and a discovery question that captures renewal dates are enough to populate them without buying an intent data product.

## Wiring it in, which is where most ICPs stop

Five operational steps turn the document into something that changes spend. Skip any of them and you have a nicer poster.

**Operationalising the ICP**

## Keeping it current

Book a 90 minute quarterly refresh. Rerun the closed won and churn queries, check win rate by tier against the thresholds, update trigger fields, and record what changed with a date. Anything older than two quarters should be treated as a hypothesis rather than a definition.

Feed the refreshed version straight into your [SaaS marketing plan template](/templates/saas-marketing-plan/) and your [SaaS marketing budget template](/templates/saas-marketing-budget/), because a tightened ICP usually implies a channel change and sometimes a budget cut. For the strategic detail behind these fields, read the longer [B2B SaaS ideal customer profile](/guides/b2b-saas-ideal-customer-profile/) guide, pair it with the matching [buyer persona](/glossary/buyer-persona/) work, and see how it connects to everything else on the [SaaS marketing](/saas-marketing/) hub.

## Frequently asked questions

### What is an ideal customer profile in SaaS?

An ICP describes the company you sell to, not the person. It combines firmographic, technographic, behavioural and trigger criteria into a definition specific enough to accept or reject an account. In SaaS it should predict retention as well as acquisition, since an account that buys and churns in seven months is a worse outcome than one you never won.

### What is the difference between an ICP and a buyer persona?

The ICP is the account, the persona is the human inside it. You target the ICP with account selection, ad audiences and routing rules. You target the persona with messaging, channel choice and content format. Both are needed, and confusing them produces campaigns aimed at a job title inside companies that could never buy the product.

### How many fields should an ICP have?

Around 20 to 25 is the working range. Fewer than 15 and the definition is too loose to route against, more than 30 and nobody maintains it. The 22 field version here splits into seven firmographic, four technographic, six behavioural and five trigger fields, with each field carrying a stated data source so it can actually be populated.

### What is a negative ICP?

A written list of the account types you refuse, with the reason. Typical entries include single seat buyers, agencies reselling under their own brand, companies below a headcount threshold, and industries whose procurement cycle exceeds your runway. It exists to stop the sales team chasing revenue that churns and to stop marketing paying for clicks that cannot convert.

### How do you score accounts against an ICP?

Assign each field a weight totalling 100, then score each field 0, 3 or 5 and normalise. Accounts above 80 route straight to an account executive, 55 to 79 go to nurture with a fit reason attached, and below 55 stay self serve only. Review the thresholds every quarter against actual win rates by band.

### How often should you update your ICP?

Trigger fields monthly, everything else quarterly, with a full rebuild annually or whenever ACV moves more than 25%. The quarterly pass reruns closed won and churn analysis and checks whether the highest scoring band still wins at the rate the model assumes. Most ICPs drift because the product changed and nobody told marketing.

### Do PLG companies need an ICP?

Yes, and often two. Product led companies need a signup level ICP for self serve fit and a separate expansion ICP describing which accounts justify a sales touch. The criteria differ sharply: self serve fit is about time to value and single user utility, while expansion fit is about seat count, budget authority and security review capacity.