Lead Scoring

Was ist Lead Scoring?

Lead Scoring ist der Unterschied zwischen "Wir kontaktieren alle 1000 Leads gleich" und "Wir fokussieren auf die 50 mit höchster Kaufwahrscheinlichkeit". Zwischen Spray-and-Pray und Precision-Targeting. Zwischen 2% Conversion und 15% Conversion. Lead Scoring ist das Filter-System, das aus Masse Qualität macht.

— Vertriebswikinger Glossar

Lead Scoring entstand in den frühen 2000ern mit Marketing-Automation-Plattformen (Marketo, Eloqua, Pardot). Davor war Lead-Priorisierung "Bauchgefühl" – heute ist es Data-driven.

Dein Sales-Team kämpft mit zu vielen unqualifizierten Leads? Wir implementieren Lead-Scoring-Systeme die Conversion-Rates verdoppeln →

Lead Scoring auf einen Blick

Eigenschaft	Wert
Definition	Numerische Bewertung von Leads basierend auf Fit + Intent
Score-Range	Typisch 0-100 Punkte (oder A-F Letter-Grade)
Komponenten	Explicit-Score (Firmographics) + Implicit-Score (Behavior)
Threshold	Hot-Lead: 70+ Punkte, Warm: 40-69, Cold: <40
Update-Frequency	Real-time (bei jedem Behavior-Event)
Primary-Use	Lead-Priorisierung für SDRs/AEs
Ownership	Marketing-Ops + Sales-Ops (gemeinsam)
Tools	HubSpot, Marketo, Salesforce, 6sense, MadKudu

Lead Scoring vs. Lead Qualification

Kriterium	Lead Scoring	Lead Qualification
Timing	Pre-Contact (automatisch)	During-Contact (manuell)
Method	Algorithmic (Data-driven)	Conversational (BANT, Discovery)
Output	Numerischer Score (0-100)	Binary (Qualified/Disqualified)
Scalability	High (tausende Leads)	Low (requires Human-Time)
Accuracy	70-85% (Prediction)	90%+ (Direct-Validation)
Purpose	Priorisierung (Wer zuerst?)	Validation (Wert es sich überhaupt?)

Die Wahrheit: Lead Scoring ist Pre-Filter – Lead Qualifizierung ist Final-Validation

Warum Lead Scoring kritisch ist

Problem ohne Lead Scoring

Scenario: 1000 Leads in CRM SDR-Approach ohne Scoring:

• Kontaktiere Leads in chronologischer Reihenfolge (First-In-First-Out)
• Oder random Pick

Result:

• 50% der Leads sind komplett unqualified (Student, Competitor, Wrong-Industry)
• 30% sind okay-Fit, aber kein Intent (browsed Website 1x, no Follow-Up)
• 15% sind Good-Fit, Low-Intent
• 5% sind Perfect-Fit, High-Intent (kaufbereit!)

SDR verschwendet 80% der Zeit auf die falschen 95% der Leads

Conversion-Rate: 2% (20 von 1000 werden zu SQL)

---

Lösung mit Lead Scoring

Same 1000 Leads, aber:

• Lead-Scoring-Model bewertet jeden Lead (0-100 Punkte)
• Top-50-Leads (80-100 Punkte) = Perfect-Fit + High-Intent
• SDR kontaktiert diese 50 zuerst

Result:

• 50 Leads mit 40% Conversion = 20 SQLs
• Gleiche Output, aber 95% weniger Zeitaufwand
• SDR kann jetzt 20× mehr Leads in gleicher Zeit bearbeiten (1000 → 50 = 20x Effizienz)

Conversion-Rate (Top-50): 40% (vs. 2% ohne Scoring)

Impact: 20x Effizienz-Steigerung

---

Business-Impact

1. Höhere Conversion-Rates:

• Fokus auf High-Score-Leads → 3-5x höhere SQL-Conversion

2. Kürzere Sales-Cycles:

• High-Intent-Leads sind kaufbereit → schneller Close

3. Bessere SDR-Productivity:

• Weniger Zeit auf Junk-Leads → mehr Zeit auf Real-Opportunities

4. Predictable Pipeline:

• "Wir haben 200 Hot-Leads (80+ Score) → erwarten 80 SQLs → forecaste 24 Deals"

5. Marketing-Accountability:

• "Marketing generiert Leads mit AVG-Score 55 → vs. Last-Quarter: 45 → Improvement!"

Die Komponenten eines Lead-Scoring-Models

Komponente 1: Explicit Scoring (Fit)

Definition: Demografische/Firmografische Attribute – "Passt Lead zu unserem ICP?"

Beispiel-Attributes: Company-Size (Employees):

• 1-10: -10 Punkte (zu klein)
• 11-50: 0 Punkte
• 51-200: +10 Punkte (Sweet-Spot)
• 201-1000: +15 Punkte (Ideal)
• 1000+: +5 Punkte (zu groß, komplex)

Industry:

• FinTech: +20 Punkte (Target-Vertical)
• Healthcare: +15 Punkte
• Retail: +10 Punkte
• Education: 0 Punkte (Not-Ideal)
• Government: -10 Punkte (zu lange Sales-Cycles)

Job-Title:

• CEO, VP, Director: +15 Punkte (Decision-Maker)
• Manager: +10 Punkte
• Individual-Contributor: +5 Punkte
• Student, Intern: -20 Punkte (unqualified)

Geography:

• DACH-Region: +15 Punkte (Home-Market)
• EU: +10 Punkte
• USA: +5 Punkte
• Rest-of-World: 0 Punkte

Revenue (Company-Revenue):

• <1M €: -5 Punkte
• 1-10M €: +10 Punkte
• 10-50M €: +15 Punkte (Sweet-Spot)
• 50M+ €: +10 Punkte

Technology-Stack:

• Uses Salesforce: +10 Punkte (Integration-Fit)
• Uses HubSpot: +10 Punkte
• Uses Competitor-Tool: -5 Punkte (Switching-Friction)

Funding-Stage (for Startups):

• Series B+: +15 Punkte (Budget vorhanden)
• Series A: +10 Punkte
• Seed: +5 Punkte
• Pre-Seed: -5 Punkte (zu früh)

---

Explicit-Score-Calculation: Example-Lead:

• Company-Size: 150 Employees → +10
• Industry: FinTech → +20
• Job-Title: VP Sales → +15
• Geography: Germany → +15
• Revenue: 20M € → +15
• Tech-Stack: Salesforce → +10

Total-Explicit-Score: 85 Punkte (Excellent-Fit!)

---

Komponente 2: Implicit Scoring (Intent/Behavior)

Definition: Verhaltens-Signale – "Zeigt Lead Kaufinteresse?"

Beispiel-Behaviors: Website-Engagement:

• Visited Homepage: +2 Punkte
• Visited Pricing-Page: +10 Punkte (High-Intent!)
• Visited Case-Studies: +5 Punkte
• Visited Careers-Page: -5 Punkte (Recruiting, nicht Buying)
• 5+ Page-Views: +5 Punkte (Deep-Engagement)
• Spent 10+ Minutes: +5 Punkte

Content-Downloads:

• Downloaded Whitepaper: +5 Punkte
• Downloaded Product-Guide: +10 Punkte
• Downloaded ROI-Calculator: +15 Punkte (sehr kaufnah)

Email-Engagement:

• Opened Email: +2 Punkte
• Clicked Link: +5 Punkte
• Replied to Email: +15 Punkte (Direct-Engagement!)
• Unsubscribed: -20 Punkte

Event-Participation:

• Attended Webinar: +10 Punkte
• Attended Conference-Booth: +15 Punkte
• Requested Demo: +25 Punkte (Hot-Lead!)

Product-Interaction:

• Started Free-Trial: +20 Punkte
• Logged-In 3+ Times: +10 Punkte
• Invited Team-Members: +15 Punkte (Serious-Intent)
• Used Key-Feature: +10 Punkte

Sales-Interaction:

• Responded to Outreach: +15 Punkte
• Scheduled Meeting: +25 Punkte
• No-Show to Meeting: -15 Punkte
• Rescheduled Meeting: +5 Punkte (still engaged)

Time-Decay:

• Recent-Activity (last 7 Days): 1.0× Multiplier
• 8-30 Days ago: 0.75× Multiplier
• 31-90 Days ago: 0.5× Multiplier
• 90+ Days ago: 0.25× Multiplier (Interest cooled)

Frequency-Bonus:

• 1 Website-Visit: +2 Punkte
• 5 Visits in 7 Days: +2 × 5 + 5 Bonus = +15 Punkte (Multiple-Touchpoints = stronger Intent)

---

Implicit-Score-Calculation: Example-Lead:

• Visited Pricing-Page (Yesterday): +10
• Downloaded ROI-Calculator (3 Days ago): +15
• Opened 3 Emails (Last 2 Weeks): +6
• Requested Demo (Today): +25
• 7 Page-Views (Last Week): +7

Total-Implicit-Score: 63 Punkte (High-Intent!)

---

Komponente 3: Combined Score

Total-Lead-Score = (Explicit-Score × Weight-A) + (Implicit-Score × Weight-B) Typical-Weighting:

• Option 1 (Equal-Weight): 50% Fit + 50% Intent
• Option 2 (Fit-Heavy): 60% Fit + 40% Intent (Enterprise-Sales, wo ICP critical)
• Option 3 (Intent-Heavy): 40% Fit + 60% Intent (Product-Led-Growth, wo Usage-Signals key)

Example (Equal-Weight):

• Explicit: 85 Punkte
• Implicit: 63 Punkte
• Combined: (85 × 0.5) + (63 × 0.5) = 42.5 + 31.5 = 74 Punkte (Hot-Lead!)

---

Lead-Grading: Score + Grade

Many-Companies nutzen 2-Dimensional-System:

Dimension 1: Score (0-100) = Intent
Dimension 2: Grade (A-F) = Fit

Matrix:

Fit-Grade	Score 80-100	Score 60-79	Score 40-59	Score <40
A (Excellent-Fit)	A1 (Top-Priority!)	A2 (High-Priority)	A3 (Nurture)	A4 (Long-Term)
B (Good-Fit)	B1 (High-Priority)	B2 (Medium-Priority)	B3 (Nurture)	B4 (Low-Priority)
C (Okay-Fit)	C1 (Medium-Priority)	C2 (Low-Priority)	C3 (Nurture)	C4 (Ignore)
D-F (Poor-Fit)	D1 (Disqualify)	D2 (Disqualify)	D3 (Disqualify)	D4 (Disqualify)

Action-Mapping:

A1-Leads (Perfect-Fit + High-Intent): SDR calls within 5 Minutes (Strike-while-Hot!)

A2/B1 (Good-Fit + Medium/High-Intent): SDR calls within 24 Hours

A3/B2 (Good-Fit + Low-Intent): Add to Nurture-Cadence

C1-C3: Marketing-Nurture (not Sales-Ready yet)

D-F (Any-Score): Disqualify (Wrong-ICP) oder Remove (Competitor, Student, etc.)

Die 3 Lead-Scoring-Methoden

Methode 1: Manual/Rule-Based Scoring

Wie es funktioniert: Step 1: Define Rules

• If Company-Size = 100-500 → +15 Punkte
• If Industry = FinTech → +20 Punkte
• If Downloaded Whitepaper → +5 Punkte
• If Visited Pricing → +10 Punkte

Step 2: Implement in CRM/Marketing-Automation

• HubSpot-Workflow: "When Contact-Property 'Industry' = 'FinTech', Add 20 to Lead-Score"
• Salesforce-Process-Builder: Same-Logic

Step 3: Score updates in real-time

• Lead fills Form mit Industry "FinTech" → Score +20
• Lead visits Pricing → Score +10
• New-Total-Score: 30

---

Pros:

• Simple (no AI needed)
• Transparent (everyone understands Rules)
• Controllable (manually adjustable)

Cons:

• Manual-Maintenance (Rules müssen constantly updated werden)
• Limited-Accuracy (Can't capture complex Patterns)
• Static (No Learning from Outcomes)

Best-for: Small-Companies (10-50 Reps), Limited-Data (<1000 Leads/Month)

Accuracy: 70-75% (okay, but not great)

---

Methode 2: Predictive/AI-Powered Scoring

Wie es funktioniert: Step 1: Historical-Data-Analysis

• ML-Model analyzed 10,000+ past Leads
• Identified Patterns: "Leads mit X-Attributes + Y-Behaviors haben 40% Conversion-Rate"

Step 2: Model-Training

• Algorithm learned:

- "FinTech-Leads mit >100 Employees, visited Pricing 2x, downloaded Case-Study = 60% Conversion" - "Retail-Leads mit <50 Employees, 1 Page-View = 2% Conversion" Step 3: Score-Prediction

• New-Lead kommt rein
• Model analyzed 100+ Attributes + Behaviors
• Output: Predictive-Score: 82% Likelihood-to-Convert → mapped to 0-100-Scale = 82 Punkte

Step 4: Continuous-Learning

• Model re-trained monatlich
• Learns from new Closed-Won/Lost-Leads
• Accuracy improves over Time

---

Pros:

• Higher-Accuracy (85-90% vs. 70% Manual)
• Discovers-Hidden-Patterns (things Humans miss)
• Self-Improving (Gets better with more Data)
• Handles-Complexity (100+ Variables)

Cons:

• Requires-Data (min. 1000+ Closed-Leads for Training)
• Blackbox (Hard to explain "Why Score X?")
• Expensive (Tools: 6sense, MadKudu, Infer = 50k+ €/year)
• Setup-Time (3-6 Months to Train + Validate)

Best-for: Scale-Ups/Enterprises (100+ Reps), High-Lead-Volume (>5000 Leads/Month), Mature-Data

Accuracy: 85-90%

---

Predictive-Scoring-Tools:

MadKudu: AI-Scoring for B2B-SaaS

• Integrates mit CRM
• Real-time-Scoring
• Fit + Intent-Score

6sense: Intent-Data + Predictive-Scoring

• Account-Level-Scoring (Account-Based Selling)
• Intent-Signals from Web-Behavior

Infer (acquired by ZoomInfo): Predictive-Lead-Scoring
Leadspace: AI-Powered B2B-Data + Scoring

---

Methode 3: Hybrid (Manual + Predictive)

Wie es funktioniert: Combine Best-of-Both: Base-Score (Manual-Rules):

• Explicit-Attributes (Firmographics) → Manual-Rules
• Clear-Thresholds (Company-Size, Industry, etc.)
• Transparent + Explainable

Intent-Score (Predictive-AI):

• Behavioral-Patterns → AI-Model
• Discovers-Complex-Signals
• Continuously-Optimized

Total-Score = Base-Score (40%) + Intent-Score (60%)

---

Example:

Lead:

• Company: 200 Employees, FinTech, Germany, 15M € Revenue
• Base-Score (Manual): 75 Punkte (Excellent-Fit)
• Behavior: Visited Pricing 3x, Downloaded 2 Case-Studies, Opened 5 Emails, Attended Webinar
• Intent-Score (AI-Predicted): 88% Likelihood → 88 Punkte

Total-Score: (75 × 0.4) + (88 × 0.6) = 30 + 52.8 = 82.8 Punkte (Hot-Lead!)

---

Pros:

• Best-Accuracy (combines Transparency + ML-Power)
• Explainable (Base-Score ist clear, Intent-Score ist AI-optimized)
• Flexible (adjust Weights based auf Business-Needs)

Cons:

• Complex-Setup (need both Rule-Engine + AI-Model)
• Requires-Resources (Data-Scientists + Marketing-Ops)

Best-for: Mid-Size-to-Large-Companies (50-500 Reps), Growing-Data

Accuracy: 80-85%

Lead-Scoring-Implementation: 6-Schritte-Prozess

Schritt 1: Define Ideal-Customer-Profile (ICP)

Wer ist euer Perfect-Customer? Workshop mit Sales + Marketing + CS: Analyze Best-Customers (Top-20-Accounts):

• Company-Size: AVG 150 Employees (Range: 100-500)
• Industry: 60% FinTech, 30% Healthcare
• Geography: 80% DACH, 20% EU
• Revenue: AVG 25M € (Range: 10-100M €)
• Tech-Stack: 90% use Salesforce oder HubSpot
• Job-Titles: 70% VP/Director-Level

Common-Patterns:

• Series-B+ Funding
• Growing (20%+ YoY)
• Sales-Team >20 Reps

Output: ICP-Definition-Document

---

Schritt 2: Identify Key-Attributes & Behaviors

Attributes (Explicit): Must-Haves:

• Company-Size
• Industry
• Job-Title
• Geography

Nice-to-Haves:

• Revenue
• Funding-Stage
• Tech-Stack
• Growth-Rate

Data-Sources:

• Form-Fills (manual Input)
• Enrichment-Tools (Clearbit, ZoomInfo)
• LinkedIn-Data
• CRM-Data

---

Behaviors (Implicit): High-Intent-Signals:

• Visited Pricing-Page
• Requested Demo
• Downloaded ROI-Calculator
• Started Trial
• Multiple-Visits (5+ in 7 Days)

Medium-Intent:

• Downloaded Whitepaper
• Attended Webinar
• Email-Engagement (Opened, Clicked)

Low-Intent:

• Homepage-Visit
• Blog-Read

Negative-Signals:

• Unsubscribed
• No-Show to Meeting
• Visited Careers-Page (Recruiting-Intent, nicht Buying)

---

Schritt 3: Assign-Point-Values

Methode: Correlation-Analysis Analyze Historical-Data: Example:

• Leads mit "Visited Pricing" have 40% Conversion-Rate
• Leads without "Visited Pricing" have 10% Conversion-Rate
• → Correlation: 4x höher → Assign high Points (+10 or +15)

Compare-Attributes:

Attribute/Behavior	Conversion-Rate	Points-Assigned
Visited Pricing	40%	+15
FinTech-Industry	35%	+20
Company-Size 100-500	30%	+15
Downloaded ROI-Calc	45%	+20
Attended Webinar	25%	+10
Email-Opened	12%	+2
Homepage-Visit	8%	+1

Rule:

• High-Correlation (>3x Baseline) → +15 to +20 Punkte
• Medium-Correlation (2-3x) → +10 to +15
• Low-Correlation (1.5-2x) → +5 to +10
• Minimal-Correlation (<1.5x) → +1 to +5

Negative-Points:

• Wrong-ICP (Student, Competitor, Too-Small) → -10 to -20
• Disengagement (Unsubscribe, No-Show) → -10 to -20

---

Schritt 4: Set-Thresholds

Define Score-Ranges: Hot-Lead (80-100 Punkte):

• Action: SDR calls within 5 Minutes
• Expected-Conversion: 40-60%

Warm-Lead (60-79 Punkte):

• Action: SDR calls within 24 Hours
• Expected-Conversion: 20-30%

Cold-Lead (40-59 Punkte):

• Action: Add to Nurture-Cadence (Email-Sequence)
• Expected-Conversion: 10-15%

Unqualified (<40 Punkte):

• Action: Long-Term-Nurture oder Disqualify
• Expected-Conversion: <5%

Calibration:

• Test Thresholds über 3 Months
• Adjust based auf Actual-Conversion-Rates
• Goal: 80+ Leads convert at 40%+, 60-79 at 20%+, etc.

---

Schritt 5: Implement in Tech-Stack

CRM + Marketing-Automation: HubSpot-Example: Create-Lead-Score-Property:

• Custom-Property: "Lead-Score" (Number, 0-100)

Build-Workflows: Workflow 1: Explicit-Scoring

• Trigger: Contact-Created oder Property-Changed
• Actions:

- If Industry = FinTech → Increase Lead-Score by 20 - If Company-Size = 100-500 → Increase by 15 - If Job-Title contains "VP" → Increase by 15 Workflow 2: Implicit-Scoring

• Trigger: Contact-Visited-Page
• Actions:

- If Page = Pricing → Increase by 15 - If Page = Case-Study → Increase by 5 Workflow 3: Time-Decay

• Trigger: Daily (Scheduled)
• Actions:

- If Last-Activity >30 Days → Decrease Score by 10% - If Last-Activity >90 Days → Decrease by 50% Workflow 4: Lead-Routing

• Trigger: Lead-Score-Changed
• Actions:

- If Score ≥80 → Assign to SDR-Queue "Hot-Leads" - If Score 60-79 → Assign to "Warm-Leads" - If Score <40 → Add to Nurture-Campaign

---

Salesforce-Example: Process-Builder oder Flow:

• Similar-Logic wie HubSpot
• Update Lead-Score-Field
• Assignment-Rules based auf Score

---

API-Integration (for Predictive-Scoring): MadKudu-Integration:

• MadKudu-API analyzed Lead-Data
• Returns-Score (0-100)
• Writes to CRM-Field "Predictive-Score"
• Updates in real-time

---

Schritt 6: Test, Measure, Iterate

Phase 1: Pilot (Month 1-2) Test mit Small-Group:

• 10 SDRs nutzen Scored-Leads
• 10 SDRs nutzen Traditional-Approach (Control-Group)

Measure:

• SDR-A (Scored-Leads): 40 SQLs from 100 Hot-Leads = 40% Conversion
• SDR-B (Traditional): 20 SQLs from 200 Random-Leads = 10% Conversion

Result: Scored-Leads convert 4x besser

---

Phase 2: Roll-Out (Month 3)

Full-Team-Adoption:

• All SDRs nutzen Lead-Score
• Prioritize 80+ Leads first

Measure-Weekly:

• Total-Hot-Leads (80+): 200/week
• Conversion-to-SQL: 80 (40%)
• vs. Previous-Quarter: 60 (15%)

Result: 33% mehr SQLs pro Week

---

Phase 3: Iterate (Month 4+)

Analyze-Misses: False-Positives (High-Score, but didn't Convert):

• Lead had 85 Punkte, aber disqualified in Discovery (No-Budget)
• → Adjust-Model: Budget-Attribute important → Add "Funding-Stage" with higher Weight

False-Negatives (Low-Score, but converted):

• Lead had 55 Punkte, aber wurde Champion und closed
• → Analyze: Was wurde übersehen? (Maybe "Replied to Email" is stronger Signal than thought)
• → Adjust-Points

Quarterly-Model-Review:

• Re-analyze Conversion-Rates
• Update-Point-Values
• Add/Remove-Attributes
• Re-calibrate-Thresholds

Continuous-Improvement: After 6 Months, Accuracy should improve 10-15%

Lead-Scoring Best Practices

Best Practice 1: Start Simple, Then Iterate

Don't build 100-Attribute-Model on Day-1: V1-Model (Month 1):

• 5 Explicit-Attributes (Company-Size, Industry, Job-Title, Geography, Revenue)
• 5 Implicit-Behaviors (Pricing-Visit, Demo-Request, Email-Click, Webinar-Attend, Content-Download)
• Total: 10 Inputs

V2-Model (Month 6):

• Add 10 more Attributes (Tech-Stack, Funding, Growth-Rate, etc.)
• Refine-Point-Values based auf Data

V3-Model (Month 12):

• Introduce-Predictive-Scoring (AI-Model)
• 50+ Inputs

Result: Accuracy steigt von 70% (V1) → 80% (V2) → 90% (V3)

---

Best Practice 2: Negative-Scoring ist wichtig

Don't only add Points – also subtract: Disqualifying-Attributes:

• Student-Email (@university.edu): -20 Punkte
• Competitor-Domain: -50 Punkte (Remove from List)
• Company-Size <10: -10 Punkte (too small)

Disengagement-Behaviors:

• Unsubscribed: -20 Punkte
• No-Show to Meeting: -15 Punkte
• Marked-Email-as-Spam: -30 Punkte

Result: Prevents wasting Time auf Junk-Leads

---

Best Practice 3: Time-Decay für Implicit-Score

Problem: Lead hatte hohe Activity vor 6 Monaten, dann nichts

Without Time-Decay:

• Lead-Score = 85 (from old Activity)
• SDR calls → Lead ist cold ("Oh, I looked at that months ago, not interested now")

With Time-Decay:

• Lead-Score = 85 → after 90 Days with no Activity → decays to 42
• SDR doesn't waste Time

Implementation:

• Implicit-Score decays 10% per Month without Activity
• Explicit-Score bleibt konstant (Firmographics ändern sich nicht schnell)

---

Best Practice 4: Separate Fit & Intent

Problem: High-Fit-Low-Intent-Lead vs. Low-Fit-High-Intent-Lead

Example: Lead A:

• Perfect-ICP (Large-Company, FinTech, VP-Level) → Fit-Score: 90
• No-Website-Activity → Intent-Score: 10
• Combined-Score: 50 (Medium)

Lead B:

• Okay-ICP (Small-Company, Adjacent-Industry) → Fit-Score: 40
• Requested-Demo, Visited-Pricing-5x → Intent-Score: 95
• Combined-Score: 67.5 (Warm)

Action-Difference: Lead A (High-Fit, Low-Intent):

• Don't call now (not ready)
• Add to Long-Term-Nurture (build Intent over Time)

Lead B (Low-Fit, High-Intent):

• Call immediately (they're ready!)
• But qualify hard in Discovery (confirm Fit)

Solution: Track both Scores separately (not just Combined)

---

Best Practice 5: Sales-Marketing-Alignment on Definition

Biggest-Failure-Point: Marketing + Sales disagree on "What is Hot-Lead?"

Marketing: "80+ Score = Hot"
Sales: "I called 10× 80+ Leads, only 2 were actually good"

Root-Cause: Misaligned-ICP oder Wrong-Point-Values

Solution: Monthly-Alignment-Meeting:

• Sales gives Feedback: "80+ Leads from Industry X didn't convert"
• Marketing adjusts Model: "Reduce Industry-X-Points from +20 to +10"
• Re-test for 30 Days
• Iterate

SLA (Service-Level-Agreement):

• Marketing delivers 100 Hot-Leads (80+) per Month
• Sales commits to contact within 24 Hours
• Sales provides Feedback on Lead-Quality (Accepted/Rejected)
• Target: 80% Acceptance-Rate (if lower, adjust Model)

---

Best Practice 6: Account-Based-Scoring (for Account-Based Selling)

B2B-Reality: Multiple-Contacts from same Account

Example:

• Contact-A (VP): Individual-Score = 75
• Contact-B (Director): Individual-Score = 65
• Contact-C (Manager): Individual-Score = 55

Account-Level-Score:

• Aggregate all Contacts
• AVG-Score = (75 + 65 + 55) / 3 = 65 ODER
• MAX-Score = 75 (highest Contact) ODER
• SUM-Engagement = Combined-Intent-Signals

Account-Scoring-Model:

• Fit: Company-Firmographics (1× per Account)
• Intent: Sum of all Contact-Behaviors
• Multi-Contact-Bonus: +10 Punkte if 3+ Contacts engaged (shows buying-committee forming)

Result: Better für Enterprise-Sales with long buying-cycles

Lead-Scoring Metriken

Input-Metriken (Model-Health)

1. Score-Distribution:

• % Leads in each Range (80+, 60-79, 40-59, <40)

Healthy-Distribution:

• 10-15% Hot (80+)
• 20-30% Warm (60-79)
• 30-40% Cold (40-59)
• 20-30% Unqualified (<40)

Red-Flag: Wenn 80% of Leads sind <40 → Model ist zu strict ODER Lead-Quality ist schlecht

---

2. Average-Lead-Score:

• AVG-Score aller Leads

Benchmark:

• 45-55 (balanced)
• If AVG <40 → Lead-Quality-Problem
• If AVG >60 → Model zu lenient (inflated Scores)

---

Output-Metriken (Business-Impact)

3. Conversion-Rate by Score-Range:

Score-Range	Leads	SQLs	Conversion-Rate
80-100	150	60	40%
60-79	300	75	25%
40-59	500	50	10%
<40	1000	20	2%

Target: Clear-Gradient (Higher-Score = Higher-Conversion)

Red-Flag: Wenn 60-79-Leads convert besser als 80-100 → Model ist mis-calibrated

---

4. Model-Accuracy:

Precision: Von allen Hot-Leads (80+), wieviele wurden wirklich SQLs?

Formel: True-Positives / (True-Positives + False-Positives)

Example:

• 100 Leads scored 80+
• 40 became SQLs (True-Positives)
• 60 didn't (False-Positives)
• Precision: 40 / (40 + 60) = 40%

Benchmark: 40-60% Precision

---

Recall: Von allen SQLs, wieviele waren Hot-Leads (80+)?

Formel: True-Positives / (True-Positives + False-Negatives)

Example:

• 50 Total-SQLs
• 40 were from 80+ Leads (True-Positives)
• 10 were from <80 Leads (False-Negatives – Model missed them)
• Recall: 40 / (40 + 10) = 80%

Benchmark: 60-80% Recall

---

5. SDR-Efficiency:

Time-to-Contact (for Hot-Leads):

• Target: <5 Minutes (Strike-while-hot!)

Connect-Rate:

• % Hot-Leads wo SDR actual Conversation hatte
• Benchmark: 30-50% (vs. 5-15% for Cold-Leads)

SQL-Conversion:

• % Hot-Leads → SQL
• Benchmark: 30-50% (vs. 2-5% unscored Leads)

---

6. Revenue-Impact: Pipeline from Hot-Leads:

• Total-Pipeline-Value from 80+-Leads

Example:

• 100 Hot-Leads → 40 SQLs → 12 Closed-Won = 12% Win-Rate
• AVG-Deal-Size: 50k €
• Revenue: 600k €

vs. Unscored-Approach:

• 1000 Random-Leads → 20 SQLs → 6 Closed-Won = 0.6% Win-Rate
• Revenue: 300k €

Result: Lead-Scoring verdoppelt Revenue bei gleichem Effort

Häufige Lead-Scoring-Fehler

Fehler 1: Too-Complex-on-Day-1

Problem: Building 50-Attribute-Model sofort

Result:

• Overwhelming
• Hard to maintain
• Inaccurate (not enough Data für alle Attributes)

Fix: Start with 5-10 Key-Attributes, iterate

---

Fehler 2: Set-It-and-Forget-It

Problem: Model gebaut, never updated

Reality: ICP changes, Buyer-Behavior changes, Product changes

Result: Model-Accuracy decays (70% → 50% over 1 year)

Fix: Quarterly-Review + Update

---

Fehler 3: Ignoring-Sales-Feedback

Problem: Marketing builds Model in Silo, Sales says "These Leads suck"

Result: Sales ignores Scores, Model ist nutzlos

Fix: Co-create Model mit Sales, iterate based auf Feedback

---

Fehler 4: No-Negative-Scoring

Problem: Alle Leads bekommen positive Points, keiner wird removed

Result: Junk-Leads (Students, Competitors) haben moderate Scores, waste SDR-Time

Fix: Aggressively disqualify mit Negative-Points

---

Fehler 5: Scoring-without-Routing

Problem: Leads have Scores, aber kein automated Routing

Result: SDRs müssen manually CRM durchsuchen für Hot-Leads

Fix: Auto-Assign Hot-Leads (80+) to SDR-Queue, Notification sent

---

Fehler 6: Not-Tracking-Outcomes

Problem: "We have Lead-Scoring" – aber kein Measurement ob es funktioniert

Fix: Track Conversion-Rates by Score-Range, calculate ROI

Lead-Scoring Tools

All-in-One (CRM + Scoring)

HubSpot: Built-In-Lead-Scoring

• Manual-Rule-Based
• Easy-Setup
• Free on Pro-Tier+

Salesforce (with Pardot): Enterprise-Grade

• Rule-Based + Predictive (Einstein-Scoring)
• Complex-Setup

Marketo: Marketing-Automation-Leader

• Advanced-Scoring-Rules
• A/B-Testing

---

Predictive-Scoring-Plattformen

MadKudu: AI-Lead-Scoring for SaaS

• Real-time-Scoring
• CRM-Integration
• Fit + Intent-Score

6sense: Account-Based-Scoring

• Intent-Data + Predictive-AI
• Best für Account-Based Selling

Infer (by ZoomInfo): Predictive-Lead-Analytics

Leadspace: B2B-Lead-Management + Scoring

---

Data-Enrichment (für Explicit-Scoring)

Clearbit: Real-time-Enrichment

• Firmographics + Technographics
• API-Integration

ZoomInfo: B2B-Contact-Data

• 100M+ Contacts
• Intent-Data

Apollo.io: All-in-One (Data + Engagement + Scoring)

---

Intent-Data (für Implicit-Scoring)

Bombora: Company-Surge-Data (wer researcht deine Kategorie?)
G2: Review-Intent (wer compared deine Competitors?)

Zusammenfassung

Lead Scoring ist ein systematisches Framework zur Bewertung von Leads basierend auf Fit (Firmographics) + Intent (Behavior) – typischerweise als 0-100-Punkte-Score. Ziel: SDRs fokussieren auf die 5-10% vielversprechendsten Leads statt Zeit auf die falschen 90% zu verschwenden.

Die 3 Scoring-Methoden:

1. Manual/Rule-Based: Simple-Rules (If Industry = X → +20 Points) → 70-75% Accuracy → Best für Small-Companies

2. Predictive/AI-Powered: ML-Model trained auf Historical-Data → 85-90% Accuracy → Best für Enterprises mit Data

3. Hybrid: Base-Score (Manual) + Intent-Score (AI) → 80-85% Accuracy → Best für Mid-Size

Die 2 Score-Components:

Explicit-Score (Fit): Company-Size, Industry, Job-Title, Geography, Revenue, Tech-Stack → "Passt Lead zu ICP?"

Implicit-Score (Intent): Website-Visits, Content-Downloads, Email-Engagement, Demo-Requests → "Zeigt Lead Kaufinteresse?"

Score-Ranges & Actions:

80-100 (Hot): Call within 5 Minutes → 40-60% Conversion

60-79 (Warm): Call within 24 Hours → 20-30% Conversion

40-59 (Cold): Nurture-Cadence → 10-15% Conversion

<40 (Unqualified): Disqualify oder Long-Term-Nurture → <5% Conversion

Die Wahrheit über Lead-Scoring:
Ohne Scoring: 2% Conversion, SDRs verschwenden 80% Zeit auf Junk-Leads. Mit Scoring: 40% Conversion auf Hot-Leads, 20x Effizienz-Steigerung, SDRs fokussieren auf Real-Opportunities. Die besten Sales-Orgs nutzen Hybrid-Models (Manual + AI), reviewen quarterly, iterieren basierend auf Sales-Feedback. Lead-Scoring ist nicht "Nice-to-Have" – es ist Critical ab 1000+ Leads/Month.

Nächster Schritt: Define ICP (Workshop mit Sales + Marketing). Identify 5 Key-Explicit-Attributes + 5 Key-Behaviors. Assign Point-Values (based auf Correlation-Analysis). Set-Thresholds (80+ = Hot). Implement in HubSpot/Salesforce (Workflows). Test for 30 Days. Measure Conversion-Rates. Iterate Quarterly. Nach 6 Monaten: 10-15% Accuracy-Improvement realistisch.