Enterprise sales software treats lead scoring like a science experiment with 47 input variables and machine learning. Service businesses do not need any of that. Five variables explain almost all the variance in whether a lead actually books. This post walks through each variable, how the score gets calculated, and how the priority queue works in practice.
Why service business lead scoring is different
B2B SaaS sales has multiple decision-makers, long evaluation cycles, and complex stakeholder mapping. Service business sales is one person calling because something is broken. The decision-maker is the caller. The evaluation is "can you come fix it." The cycle is hours to days, not months. The variables that matter are simple and few.
The five variables below explain roughly 85-90% of the variance in booking probability across HonorElevate client data spanning 200+ service businesses. Adding more variables produces minimal lift and a lot of complexity. Five is the sweet spot.
Variable 1: Source channel
Where did the lead come from? Different channels have wildly different conversion rates.
| Source | Typical conversion to booking | Score points |
|---|---|---|
| Inbound phone call | 55-70% | 40 |
| Web chat with active conversation | 30-45% | 28 |
| Web form submission | 18-28% | 18 |
| SMS-only reply to MCTB | 40-50% (already hot) | 32 |
| Cold-list outreach response | 5-12% | 5 |
| Manual entry (owner walked-in lead) | varies | 15 |
The pattern: channels where the customer initiated contact with high effort (calling, chatting live) score higher because the intent is already qualified by the channel itself.
Variable 2: Service intent
What kind of service is the customer asking about, and how urgent? Captured by the AI voice agent or web chat during the qualifying conversation.
| Intent level | Examples | Score points |
|---|---|---|
| Emergency (now or today) | "No AC, kids inside", "Water leaking", "No heat in January" | 25 |
| Same-week urgency | "AC running but not cooling well, want it looked at this week" | 18 |
| Near-term planning | "System is 18 years old, thinking about replacement this fall" | 10 |
| Information only | "Just curious what a new unit might cost someday" | 3 |
Emergencies book at 80%+. Information-only inquiries book at maybe 8%. Scoring intent correctly catches both extremes.
Variable 3: Service area match
Is the customer inside your serviceable territory?
| Area match | Score points |
|---|---|
| Core service area (no travel fee) | 15 |
| Buffer area (small travel fee) | 8 |
| Just outside (referral candidate) | 2 |
| Way outside service area | 0 |
This variable is binary in many businesses (we serve or we do not) and graduated in others (we serve, but with surcharges). The score reflects the operational reality. Leads outside the service area get auto-routed to a "polite decline" workflow regardless of how high their other scores would be.
Variable 4: Predicted ticket size
How big is this likely to be? Captured by combining service type + system age + symptom description against your trained pricing logic.
| Ticket prediction | Score points |
|---|---|
| Replacement / capital project ($8K+ HVAC) | 15 |
| Major repair ($1K-$5K) | 10 |
| Standard service call ($300-$800) | 6 |
| Maintenance / small ($50-$300) | 3 |
The score does not penalize small tickets, just weights priority appropriately. A $200 maintenance call is still a valuable customer (LTV often higher than one big-ticket repair because they keep coming back).
Variable 5: Customer history
Have we served this customer before, or do we have a relationship signal?
| History | Score points |
|---|---|
| Returning customer (2+ jobs in CRM) | 10 |
| Referred by existing customer | 7 |
| Past customer (1 job in CRM) | 6 |
| New cold lead | 3 |
Returning customers and referrals book at dramatically higher rates than cold inbounds, even controlling for the other variables.
How the math actually works
Maximum possible: 40 (source) + 25 (intent) + 15 (area) + 15 (ticket) + 10 (history) = 105. We cap at 100.
Example calculations:
Maria Hernandez (the AC emergency from prior posts)
- Source: Phone call = 40
- Intent: Emergency (no AC, kids) = 25
- Area: Saugus, core area = 15
- Ticket: Predicted $4K range (repair territory leaning major) = 10
- History: New cold (first call) = 3
- Total: 93. Priority queue. Owner alert fires.
Hypothetical Robert from Valencia
- Source: Web form submission = 18
- Intent: Same-week ("want to schedule maintenance") = 18
- Area: Valencia, core area = 15
- Ticket: Maintenance ($89 service call) = 3
- History: Returning (3rd job in 2 years) = 10
- Total: 64. Standard queue. Normal workflow.
Hypothetical Sarah from outside Lancaster
- Source: Web form = 18
- Intent: Information only ("just exploring options") = 3
- Area: Outside service area = 0
- Ticket: Unknown = 0
- History: Cold = 3
- Total: 24. Out-of-area routing. Polite decline + referral suggestion.
What happens at each score band
80-100: Priority queue
Owner SMS alert fires immediately. If the owner is available, lead routes direct to owner. AI voice agent on Dominate tier handles same-day booking with priority slots. Pipeline tagged "High-Value Prospect". Follow-up workflows on accelerated timing (24-hour quote check-in instead of 48).
50-79: Standard queue
Standard AI handling, standard pipeline progression, standard follow-up timings (48-hour quote check-in, day-before reminder, etc.). Owner sees the lead in the Monday brief.
20-49: Low-priority / nurture
AI handles qualification with lower urgency. Leads tagged "Long-Cycle Prospect" or "Future Planning". Nurture workflow triggers (monthly check-in SMS, seasonal content). Owner not alerted unless score changes.
0-19: Polite decline / archive
Out-of-area, spam, or pure information requests with no real intent. Polite-decline workflow fires. Lead archived. Owner not notified.
Reweighting for your business
The default weights are tuned for typical HVAC. Other industries may want different weights.
- Roofing: ticket prediction gets more weight (single big-ticket jobs dominate revenue). 15 → 25 points.
- Dental: service intent gets less weight (most calls are routine), customer history gets more (retention is the business model). Intent 25 → 15, history 10 → 18.
- Med spa: source channel weighting shifts (Instagram DM converts higher than HVAC's typical web form). Source distribution adjusted to reflect.
- Recurring service (cleaning, lawn): service area weighting nearly doubles (route density is everything). 15 → 25.
- Real estate: intent vs ticket weighting flips entirely (a "just exploring" can become a $1.5M listing). Whole model redesigned during onboarding.
Configuration happens during onboarding. The default is a fast start. Adjustments happen as data accumulates.
Want the scoring tuned for your business?
Free 30-minute AI audit. We review your booking history, identify what predicts your specific customer base, and configure the scoring weights specific to your industry and operations.
Book My Free AI AuditWhat lead scoring does NOT do
To set expectations:
- It does not replace judgment. Score 75 from a customer who is being rude or asking sketchy questions still deserves the owner's instinct. Score 30 from a long-term customer making a small request still deserves attention.
- It does not predict every individual lead. Aggregate accuracy is high. Individual leads can defy the score (low-score leads sometimes become huge customers, high-score leads sometimes ghost). The point is to route attention proportionally to the aggregate likelihood.
- It does not change conversion rates. Scoring is triage, not magic. The conversion improvement comes from better-allocated attention, not from the score itself.
The owner workflow actually changes
Pre-HonorElevate, most owners triage by guess. Whoever called most recently or sounded most upset got the attention. High-value leads sometimes sat in voicemail while the owner was on the phone with a $89 service-call inquiry.
Post-deploy, the priority queue surfaces the right leads first. Score 95 lead fires an alert that interrupts whatever the owner is doing. Score 60 lead sits in the standard queue. Score 25 lead nurtures itself. The owner's attention goes to the leads where attention matters most.
This is one of the operational changes that produces real, measurable revenue lift inside the first 90 days. Same calls coming in. Same owner. Different attention allocation. Different outcomes.
The bottom line
Lead scoring for service businesses does not require machine learning, complex models, or spreadsheets. Five variables, default weights, auto-calculated 0-100 score, threshold-based routing. The platform does the math the moment a lead enters. The owner sees the right leads first.
The lift comes from better attention allocation. Same leads as before. Same conversion math. Better triage produces 10-25% more closed jobs inside the first quarter for most operators.
For the pillar context, read The Complete Guide to CRM and Pipeline. For the tags and custom fields that feed the scoring inputs, read Tags and Custom Fields: How Smart Segmentation Beats Generic CRM.