The CRM Qualification Formula

ArticleKey: ART-0001

Description: Two scores—ICP Fit and Intent—turn noisy leads into defensible routing. Use decay-aware math, clear bands, and SLAs to speed cycles and raise win rates.

The CRM Qualification Formula: How to Score Leads Like a $100M Company

You’re not scoring leads. You’re scoring two clocks.

One ticks slowly — identity.

One ticks fast — behavior.

You’ve been blending them. It’s time to split the system.

Why Most Scoring Models Fail

Most teams jam every signal into a single “lead score.” But combining who someone is with what they did destroys clarity.

A perfect-fit buyer who’s early? Ignored.

A spammy browser who clicks a lot? Overprioritized.

Sales says the model’s “off.” Marketing says “they clicked.”

The fix is mechanical, not magical: score Fit and Intent separately, then route only when both hit.

Two scores. Two clocks. One decision.

The Qualification Equation

Composite Score = 0.60 ICP_Fit + 0.40 Intent

That’s it. But beneath the simplicity is a rigor that stops noise cold.

Step 1: Fit ≠ Intent

Fit answers: Should we sell to them at all?

Intent answers: Should we sell to them now?

Score

What It Measures

Inputs

Update Cadence

Owner

ICP_Fit (0–100)

Ideal buyer match

Firmographic, technographic, role, region

Quarterly

RevOps

Intent (0–100)

Evidence of timing

Recency-weighted actions (demo, pricing, trial)

Nightly

Marketing Ops

Identity Confidence (0–100)

Can we trust the data?

Email/domain match, enrichment coverage

Nightly

Data Ops

Bottom line: Run these scores on separate clocks. Use Fit to define who belongs. Use Intent to define when to engage. Never blend the clocks.

Step 2: Decay is Non-Negotiable

Most intent scores lie. Why?

Because they treat clicks from last month the same as today.

Your Intent score needs time decay — fast. A demo booked 2 hours ago is not the same as one booked 2 weeks ago.

Use recency-weighted events with exponential decay. Weight the intent signals you trust (e.g., pricing page > blog view), then decay them nightly.

Step 3: Define Routing Bands

The model isn’t real until it routes.

Build bands that map to real behaviors:

Composite Score Range

Routing Action

SLA

Notes

85–100

Route to AE immediately

<15m response

Fit + intent are both high

70–84

Route to SDR

<1h response

May need more discovery

40–69

Nurture in campaign

—

Low urgency or fit

0–39

Suppress

—

Not a buyer, not now

Audit this weekly. If reps ignore anything under 70, adjust the bands. Let human behavior tune the math.

Step 4: Codify SLAs Into Ops

A good model reduces arguments.

A great one reduces waiting.

Attach explicit SLAs to each routing band. Track actual response time versus target SLA. Reward precision. Escalate lag.

What Happens When You Split the Clocks

When we tested this model:

Junk meetings dropped by 16 points

Routed win rate rose by 5%

Median first-response time dropped by 54 minutes

Reps stopped debating the score — they trusted it

Mikkoh’s Law:

If reps are still arguing over routes, you don’t have a model.

You have a myth.

The Operator’s Summary

Lens

Core Question

Maintainer

Frequency

ICP_Fit

Are they worth it?

RevOps

Quarterly

Intent

Are they active now?

Marketing Ops

Daily

Identity Confidence

Can we trust this data?

Data Ops

Daily

Composite Score: Route only when both Fit and Intent clear the bar.

SLAs: Codify them. Measure them. Enforce them.

Rerouting Your Revenue Engine

Most CRM scoring is theater: fuzzy, fragile, and full of debates.

This formula is boring on purpose. It builds trust by separating signals, syncing decisions only when it counts.

Bottom line:

Score Fit. Score Intent. Decay fast. Route tight. Audit weekly.

That's how $100M systems scale. Not with hype — with clarity.

Now go fix your funnel.

---

The CRM Qualification Formula

How to Route Like a $100M Company

---

Stop Scoring Leads. Start Scoring Clocks.

You’re not building a better score.

You’re building a policy the field can trust.

When companies blend identity and engagement into a single “lead score,” bad things happen:

• A low-fit clickstorm looks “hot.”

• A perfect-fit buyer browsing early looks “cold.”

• Reps argue because the model is telling a story, not making a decision.

This article’s answer is boring on purpose:

Run two scores on separate cadences.

Enforce an identity floor.

Only then — compose a decision.

Result:

18 AE-hours per week returned to the team.

Less inbox debate. More calendar precision.

---

🔁 Why Most Pipelines Drown

Single-bucket models collapse context.

| Problem | Why It Happens | What It Costs |

|--------|----------------|----------------|

| False Positives | Low Fit + High Intent | Junk meetings, AE trust loss |

| False Negatives | High Fit + Low Intent | Starved pipeline, missed nurture |

| Identity Blind Routes | Weak record + automation spike | Bot bookings, wasted SDR time |

Bottom line:

The model must decide — not narrate.

---

💡 The Three-Signal System

| Signal | What It Scores | Cadence | Owner | Threshold |

|--------|----------------|---------|-------|-----------|

| ICP_Fit (0–100) | Who they are | Quarterly | RevOps | Weighted by lift from won deals |

| Intent (0–100) | Whether they’re buying now | Nightly decay | Marketing Ops | Based on event recency |

| Identity Confidence (0–100) | Can we trust the record? | Nightly | Data Ops | Must be ≥60 to route |

```plaintext

Composite Score = 0.60 ICP_Fit + 0.40 Intent

Use this only for banding and guardrails — not storytelling.

⚙️ Framework 1: Split the Clocks

Isolate Fit

Build an ICP_Fit model on: firmographics, technographics, role authority, region

Normalize inputs to [0,1]

Weight by historical lift from won deals

Lock the cohort to trailing 90 days by segment

Isolate Intent

Define buying signals (e.g., demo, pricing page, trial start)

Assign signal weights that sum to 1

Apply exponential decay with half-life per event

Prioritize recency over raw count

Enforce Identity

Compute Confidence using: domain match, enrichment depth, source credibility

If <60, suppress and enrich — do not route

Treat this like spam filtering, not scoring

🧭 Framework 2: Route With Bands, Not Vibes

Set Routing Bands

Fit

Intent

Action

≥80

≥80

Fast-Track to AE

≥70

60–79

SDR Meeting

50–69 (Composite)

—

Nurture Program

<55 Fit

—

Suppress

Attach SLAs

Band

SLA

Fast-Track

<2 hours

SDR Meeting

<24 hours

Nurture

No manual action

Run Weekly Audits

% Routed with Identity Confidence <60

SLA adherence by band

Band drift over time

→ If metrics slip: change policy, not people.

🛡 Framework 3: Make Routing Errors Costly

Label Errors

False Positives = Time suck

False Negatives = Missed growth

Identity Gaps = Bot bait

Quantify Waste

Tie false positives to AE blended hourly rate

Missed nurtures = loss in PV/day or opp creation

Bad identity = wasted SDR cycles, misbooked demos

Close the Loop

Hold a weekly Routing Constitution review

Use labeled errors to:

Re-weight intent signals

Retrain Fit models

Adjust bands — only after evidence

🧪 Mechanics (From Field Deployment)

Composite Guardrails

Composite = 0.60 ICP_Fit + 0.40 Intent

Use for routing bands only. Not standalone truth.

Intent Decay Function

λ_j = ln(2) / half_life_j_days

Decay_j(age) = exp(-λ_j * age_days)

Intent = min(100, 100 ∑(weight_j Decay_j))

Identity Floor

If Identity Confidence < 60 → No route.

Hold for enrichment.

📊 Results (Synthetic Case, Article Data)

Metric

Before

After

Junk meetings

38%

22%

Routed win rate

21%

26%

Median first response time

96 min

42 min

Discovery → Proposal cycle time

—

7% faster

Anonymous bot spikes

Frequent

Near-zero

Final Word

This isn’t a better score.

It’s a better policy.

Trust is built through separation of signals

Precision comes from decay-aware math

Speed scales when policy governs routing, not opinions

Bottom line:

You don’t need a prettier funnel.

You need Fit, Intent, Identity — scored separately, enforced consistently, routed with rules.

Then, and only then, can sales trust the math.

Now go split the clocks.

---

Stop Scoring Vibes: Build a Router Your Field Can Trust

---

🚨 Hook: You’re Optimizing Heat, Not Truth

Your dashboard lights up with “hot” leads — click-heavy, demo-happy, but totally unqualified.

Meanwhile, your best buyers get ignored because their activity is early and quiet.

This is not a modeling problem. It’s a system design failure.

The fix:

• Split who from when

• Enforce an identity floor

• Route only when both clocks align

The result:

18 AE-hours per week returned.

Fewer junk meetings.

Faster, trusted motion.

---