Ideal Customer Profile Scoring Rubric — Examples, Framework, and Template (2026)

What an ICP scoring rubric is (and why generic ICPs don't convert)

An ICP scoring rubric is the operational version of your ideal customer profile. A generic ICP — "mid-market B2B SaaS in North America with 50–500 employees" — is descriptive and lives in slide decks. A scoring rubric assigns numerical weights to each attribute, sums them into a 0–100 score per prospect, and ties that score to routing rules, SLAs, and forecast categories. The descriptive ICP tells marketing who to target; the rubric tells the AE which prospect to call first this morning.

The reason generic ICPs do not convert is straightforward: they cannot be acted on at the deal level. When every prospect "kind of" matches the ICP, every rep claims every deal is qualified. The forecast inflates, the SDR team chases mediocre leads, and the CFO loses trust in the pipeline number. A rubric breaks the tie by producing a single number — 73, 41, 92 — that a manager can audit. The rubric is what actually moves pipeline math.

The bar for a working rubric is concrete: top-quartile-scored deals should close at 2–3x the win rate of the bottom quartile. If your draft rubric does not produce that spread when backfit against your last 90 closed deals, the weights are wrong. You do not have an ICP scoring problem — you have a calibration problem. Most teams ship a rubric that scores almost everyone in the 50–70 band, which means the rubric is measuring noise rather than fit. The fix is harsher weights on the 2–3 signals that actually discriminate won from lost.

One framing before we go deeper: an ICP scoring rubric is not a methodology. It does not tell reps how to run discovery, how to handle objections, or how to qualify pain. It tells reps which prospects deserve discovery in the first place. Pair the rubric with a discovery framework like MEDDPICC and a structured discovery call playbook, and the two compound — the rubric routes the right deals to the right reps; MEDDPICC then qualifies whether those deals are real.

The 4-dimension ICP scoring framework: firmographic, behavioral, intent, technographic

The four-dimension model maps to the four most predictive signal categories in our paired analysis of 47 closed-won vs 47 closed-lost B2B sales calls. No single dimension predicts win rate on its own. The combined score does — and the combination matters more than any single dimension being perfect.

The four dimensions are intentionally orthogonal — they capture different layers of "fit." Firmographic is who they are; behavioral is what they have done with you; intent is what they are doing with the wider market; technographic is whether the implementation will actually work. A prospect that scores perfectly on firmographic but zero on intent is a passive fit — interesting on paper, not actively buying. A prospect scoring high on intent but low on technographic is actively buying but might churn in month 4 because the stack does not support your product. Both patterns need to be visible in the score, which is why all four dimensions belong in the rubric.

Firmographic scoring criteria (industry, size, revenue, geography, team structure)

Firmographic is the foundation. It answers whether the prospect is structurally capable of being a customer — large enough to need your product, small enough to afford a self-serve motion, in a geography you can support, in an industry your case studies speak to. Below is the standard firmographic rubric we recommend before calibration.

Industry (0–15 points)

• 15 — Industry exactly matches a documented case-study win (e.g., B2B SaaS for a CI tool).
• 10 — Adjacent industry where you have at least 2 reference customers.
• 5 — Industry where you have a single reference customer.
• 0 — Industry with no reference customers; sale becomes a research project.

Headcount (0–15 points)

• 15 — Headcount sits in the dead center of your sweet-spot band.
• 10 — Headcount inside the band but at the edges.
• 5 — Headcount one band outside (undersized or oversized).
• 0 — Headcount two bands outside; do not pursue.

Revenue band (0–10 points)

• 10 — Revenue in the band where ACV is defensible.
• 5 — Revenue inside the band but bottom quartile.
• 0 — Revenue too low (no budget) or too high (you become a vendor of last resort).

Geography (0–5 points)

• 5 — Primary supported geography with local support hours and tax/legal infrastructure.
• 3 — Secondary supported geography.
• 0 — Unsupported geography requiring exception approvals.

Team structure (0–5 points)

• 5 — Dedicated buying team for your category exists (e.g., a RevOps function for a sales tool).
• 3 — Team responsible for the category exists informally.
• 0 — No team owns the category; sale requires creating the function before selling the product.

Firmographic total: 50 points. Scale to 40% of the total score by multiplying by 0.8. Firmographic data is the cheapest to capture (Clearbit, Apollo, ZoomInfo, LinkedIn Sales Navigator) and the slowest to change — so it is the right foundation but rarely the dimension that decides the outcome. Reps who lean entirely on firmographic miss the fact that two identically firmographic prospects can be at very different buying stages.

Behavioral scoring criteria (engagement signals, content consumption, demo requests)

Behavioral signals capture real interest from real people. They are the fastest-changing dimension and the most direct evidence that someone at the prospect company is moving. A prospect with a perfect firmographic score but zero behavioral signal is a marketing target, not a sales target.

Demo or trial action (0–10 points)

• 10 — Demo booked or trial started in the last 14 days.
• 7 — Pricing-page or ROI-calculator action in the last 14 days.
• 3 — Older demo/trial activity (15–60 days) without follow-up.
• 0 — No demo or trial activity recorded.

Content consumption depth (0–8 points)

• 8 — Multi-piece consumption across 3+ topic clusters (e.g., pricing + use-case + integration docs).
• 5 — Multi-page session on a single high-intent topic.
• 2 — Single landing-page visit.
• 0 — No tracked content consumption.

Multi-stakeholder engagement (0–7 points)

• 7 — Three or more distinct contacts from the same account engaging in 14 days.
• 4 — Two contacts engaging.
• 2 — Single contact, recurring sessions.
• 0 — Anonymous traffic only.

Behavioral total: 25 points. Already at the 25% weighting — no scaling needed. Behavioral signal refreshes weekly at minimum and ideally in real time. A score that updates monthly is functionally a marketing report, not a sales tool. If you only have monthly behavioral data, weight behavioral lower (15%) and bump firmographic up until you can capture engagement faster.

Intent scoring criteria (search behavior, comparison-shopping signals)

Intent is the most under-weighted dimension in most rubrics — and the one that best discriminates "actively buying" from "passively curious." Intent signals come from third-party data providers (G2 Buyer Intent, Bombora, 6sense, Clearbit Reveal, Demandbase) and from inferred search behavior on your own site (comparison pages, "alternatives" queries, competitor-bait pages).

G2 / review-site category visits (0–8 points)

• 8 — Account has researched your G2 category 3+ times in 30 days.
• 5 — Researched your category once or twice in 30 days.
• 2 — Adjacent-category research.
• 0 — No category research.

Comparison-shopping signals (0–7 points)

• 7 — Visited multiple "vs" pages (yours and at least one competitor's) in 14 days.
• 4 — Visited a single comparison page.
• 2 — Visited a generic alternatives page.
• 0 — No comparison-shopping activity.

Bombora / third-party intent topics (0–5 points)

• 5 — Spiking on 2+ topics relevant to your category in the last 14 days.
• 3 — Spiking on a single relevant topic.
• 0 — No relevant intent topics spiking.

Intent total: 20 points. Already at the 20% weighting. The intent dimension is the cleanest signal of "this account is in-market now" — which is exactly why outbound teams should weight it higher (30%+). If you do not have G2 or Bombora data, proxy with inferred intent: visits to your /pricing, /alternatives, and buyer-intent signal pages inside a 14-day window are reasonable substitutes until you can add a paid intent feed.

Technographic scoring criteria (tech stack, integrations, buying maturity)

Technographic captures whether the deal is feasible to win and feasible to onboard. A prospect running a competing product is signalling category-awareness — sometimes a positive (the buyer understands the value) and sometimes a negative (the buyer has a switching cost). A prospect missing a key integration may simply be uninstallable.

Competing tool in production (0–6 points)

• 6 — Running a directly competing tool with a known unhappiness trigger (e.g., recent leadership change at the vendor).
• 4 — Running a competing tool with no known unhappiness trigger (category-aware but switching-cost-bound).
• 2 — Running a category-adjacent tool (e.g., a basic recorder rather than a CI platform).
• 0 — No category tooling — sale requires creating the category internally before selling the product.

Key integration present (0–5 points)

• 5 — Has the integration that powers your core flow (e.g., Salesforce for a CI tool).
• 3 — Has a partial integration that works after configuration.
• 0 — Missing the integration; sale becomes a "rip and replace" conversation.

Buying-process maturity (0–4 points)

• 4 — Has a procurement function, has a standard MSA template, has a documented vendor onboarding process.
• 2 — Has informal procurement, MSA negotiated case-by-case.
• 0 — First-time vendor with no internal buying process; deal becomes an internal-change-management project.

Technographic total: 15 points. Already at the 15% weighting. Technographic data is the trickiest to capture cleanly — most teams pull it from BuiltWith, Clearbit Reveal, Datanyze, or HG Insights, and the data quality varies by tool category. If technographic coverage is poor for your category, set the weight to 10% and redistribute the missing 5% to behavioral until you can improve coverage.

The 0–100 ICP scoring model: weighting and thresholds

With the four dimensions summed, every prospect lands at a score between 0 and 100. The score is the rubric — not the dimension breakdown, not the rep's gut, not the stage in the CRM. Use the score to gate routing rules, SLAs, and forecast categories.

Default weighting

• Firmographic — 40% (50 raw points × 0.8)
• Behavioral — 25% (25 raw points × 1.0)
• Intent — 20% (20 raw points × 1.0)
• Technographic — 15% (15 raw points × 1.0)

Tier thresholds

• Tier-A (85–100): Call within 5 minutes. AE-owned, named-account treatment, founder may shadow the first call.
• Tier-B (60–84): Call within 24 hours. SDR-led discovery, AE-attended demo.
• Tier-C (40–59): Nurture sequence. Drip campaign and re-score in 30 days.
• Tier-D (under 40): Do not pursue. Marketing-suppress and document the disqualification reason.

Recalibration cadence

Recalibrate weights every 90 days. The process: pull every closed-won and closed-lost deal from the last 90 days, score each against the current rubric, calculate the win rate inside each tier, and shift weights toward whichever dimension best discriminates won from lost in your data. Do not change the dimensions themselves — only the weights within them. Stability of dimensions is what lets you compare quarters honestly.

ICP scoring rubric examples (3 worked examples — SaaS, services, marketplace)

Below are three paste-ready rubrics, each calibrated to a different B2B motion. Copy whichever is closest to your business and edit the weights, not the structure. Each example uses the same four dimensions but tunes the within-dimension signal weights to the motion's reality.

Example #1 — Mid-market SaaS (Nimitai's own rubric)

FIRMOGRAPHIC (40 pts max)
  Industry match (B2B SaaS sweet spot)        0–15
  Headcount (50–500 reps)                     0–15
  Revenue band ($5M–$100M ARR)                0–10

BEHAVIORAL (25 pts max)
  Demo booked or pricing visit (14d)          0–15
  Multi-piece content consumption             0–5
  3+ contacts engaging from account           0–5

INTENT (20 pts max)
  G2 conversation-intelligence category       0–10
  Visited a "vs Gong" or "vs Fathom" page     0–10

TECHNOGRAPHIC (15 pts max)
  Salesforce or HubSpot in production         0–5
  Running a competing CI tool (switch trigger)0–6
  Mature buying process / procurement         0–4

TOTAL: __ / 100
Tier:  [A 85+] [B 60-84] [C 40-59] [D <40]

Example #2 — Enterprise professional services

FIRMOGRAPHIC (50 pts max — services tilts firmographic)
  Industry match (regulated industries)       0–20
  Headcount (1,000+ employees)                0–15
  Revenue band ($250M+ ARR)                   0–15

BEHAVIORAL (20 pts max)
  RFP issued or shortlist invitation          0–12
  Multi-stakeholder site engagement           0–8

INTENT (15 pts max)
  Compliance/audit research signal            0–8
  Recent incident or audit failure (PR)       0–7

TECHNOGRAPHIC (15 pts max)
  Current incumbent firm engaged              0–6
  SOC 2 / ISO 27001 buying process            0–5
  Procurement function present                0–4

TOTAL: __ / 100
Tier:  [A 80+] [B 55-79] [C 35-54] [D <35]

Example #3 — Two-sided marketplace

FIRMOGRAPHIC (30 pts max — marketplace tilts behavioral)
  Geography (active liquidity region)         0–15
  Side of marketplace (supply or demand)      0–10
  Account category match                      0–5

BEHAVIORAL (35 pts max)
  Activation event (first listing/order)      0–15
  Repeat usage in 30 days                     0–10
  Multi-contact account engagement            0–10

INTENT (20 pts max)
  Competitor marketplace research             0–10
  Category-spike search behavior              0–10

TECHNOGRAPHIC (15 pts max)
  API integration capability                  0–7
  Payment infrastructure compatibility        0–5
  Settlement and tax flow readiness           0–3

TOTAL: __ / 100
Tier:  [A 75+] [B 50-74] [C 30-49] [D <30]

Three patterns to notice across the examples: services tilts firmographic (because deal economics are dominated by who the buyer is); SaaS balances all four (because behavioral and intent are unusually rich); marketplace tilts behavioral (because the activation event matters more than the firmographic story). The thresholds also shift — services uses Tier-A at 80+ and marketplace at 75+ because the underlying score distributions differ. Calibrate thresholds to your distribution; do not import them blind.

How AI auto-scores ICP fit from call transcripts (Nimitai angle)

The single biggest reason ICP rubrics fail in production is the same reason MEDDPICC fails: nobody updates the fields. Reps will not fill in 12 enrichment fields after every call. Marketing will not maintain behavioral scores across CRM and product analytics. RevOps will not chase down technographic refreshes. The rubric exists in Salesforce as a mostly-empty formula field, and the score it produces is half-noise.

Conversation intelligence platforms like Nimitai close the gap by listening to every discovery call and extracting ICP signals directly from buyer language. The platform tags mentioned headcount, named industry, stated tech stack, articulated pain, budget posture, and committee composition — and pipes those tags into the corresponding scoring fields automatically. The rubric updates from the call rather than from the rep's memory.

The result is an ICP score that updates after every meaningful call, without anyone touching Salesforce. Managers see real scores in real time rather than stale ones reps backfill at quarter-end. Pair this with Nimitai's AI sales researcher for pre-call account intelligence and AI sales meeting prep for the in-call coaching layer, and the ICP rubric becomes a living document instead of a quarterly housekeeping ritual.

Common ICP scoring rubric mistakes (over-weighting firmographics, ignoring intent)

Most rubric failures are not failures of the framework — they are failures of calibration. Below are the eight mistakes we see most often across teams rolling out ICP scoring.

How to roll out ICP scoring across a sales team in 30 days

A realistic ICP-rubric rollout fits inside 30 days. Beyond that you lose momentum, and the rubric becomes a "RevOps project" that never ships. The four-week plan below is the one we have seen work most reliably.

Week 1 — Calibration data

Pull every closed-won and closed-lost opportunity from the last 90 days. Enrich each with firmographic, behavioral, intent, and technographic data from your existing tools. This is your calibration dataset. Without it, the rubric is opinion; with it, the rubric is evidence.

Week 2 — Draft rubric and backfit

Draft the four-dimension rubric using the 40/25/20/15 default split. Score every deal in the calibration dataset against the draft. Calculate win rate inside each tier — A, B, C, D. If the top tier wins at 2x+ the bottom tier, the rubric is calibrated. If not, shift the weights toward whichever dimension best discriminates won from lost and re-score.

Week 3 — CRM implementation

Build the rubric as a single formula field on the Account (or Lead/Contact, depending on object model) inside Salesforce or HubSpot. Add the dimension-level sub-scores as individual fields so reps can see the breakdown. Build a tier-routing automation that sets account owner and SLA based on tier. Build a dashboard showing tier distribution across pipeline and tier-wise win rate.

Week 4 — Train and enforce

Train every AE and SDR on what the rubric measures, how to read the breakdown, and how to challenge a score that feels wrong (because sometimes the data is wrong, not the rep's intuition). Enforce the routing rules: Tier-A gets a 5-minute SLA; Tier-C goes to nurture; Tier-D is suppressed. Hold a weekly tier-distribution review for the first 8 weeks to catch drift early. Pair the rollout with AI-driven sales performance tracking so the rubric and the rep coaching layer evolve together.

Frequently asked questions about ICP scoring rubrics

What is an ICP scoring rubric?

An ICP scoring rubric is a structured 0–100 model that grades every prospect against your ideal customer profile across four dimensions — firmographic, behavioral, intent, and technographic — with weights calibrated to your historical win-rate data. It converts qualitative "good fit" judgement into an auditable number that ties to routing, SLAs, and forecast categories.

How is an ICP scoring rubric different from a generic ICP definition?

A generic ICP is descriptive ("mid-market B2B SaaS, 50–500 employees, North America") and lives in a slide deck. A scoring rubric is operational: it assigns numerical weights, produces a per-prospect score, and ties that score to specific routing and forecasting actions. The generic ICP tells marketing who to target; the rubric tells sales which prospect to call first.

What are the four dimensions of an ICP scoring rubric?

Firmographic (who the company is), behavioral (what they have done), intent (what they are researching in the wider market), and technographic (what stack they run). The four dimensions are orthogonal and together explain almost all of the win-rate variance we observed in our 47-paired closed-deals analysis.

How do I weight the four dimensions?

Start with 40% firmographic, 25% behavioral, 20% intent, 15% technographic. Recalibrate after 90 days against a fresh closed-deal cohort. Outbound-heavy teams should bump intent to 30%+; PLG-assisted teams should bump behavioral to 35%+; enterprise services should bump firmographic to 50%+.

What are good ICP scoring rubric examples for SaaS companies?

A mid-market SaaS rubric typically scores firmographics 0–40, behavior 0–25, intent 0–20, technographic 0–15. Tier-A: 85+; Tier-B: 60–84; Tier-C: 40–59; Tier-D: under 40. The Nimitai rubric in the worked examples section above is a complete copy-paste template for B2B SaaS teams selling above $25K ACV.

How can AI auto-score ICP fit from sales call transcripts?

Conversation intelligence platforms like Nimitai listen to every discovery call, extract self-reported headcount, industry, stack, and pain signals directly from buyer language, and write those into the corresponding rubric fields in Salesforce or HubSpot. The score updates from the call rather than from rep data entry, which is the only sustainable way to keep an ICP rubric alive past 90 days.