Real Estate Intelligence Series | Module 2 of 6

Lead scoring:
spotting real buyers instead of chasing everyone

Of around 13 inquiries per property, one becomes a serious buyer. Your team spends most of its time on the other 12. A machine-learning model identifies from search behavior, contact patterns, and financing signals whom you should open the door for first.

Model scenario: fictional brokerage

This module is part of a six-part model calculation about a fictional mid-sized brokerage (32 sales agents, 450 transactions per year). All figures, datasets, model metrics, and euro amounts are entirely fictional. They illustrate the methodology and the kind of effect achievable. They are not results of real clients and not a performance promise. What values are achievable on your real data is established only in a pilot project.

01 The problem: every viewing costs 2.5 hours

Travel, preparation, viewing, follow-up. And over 80% of them lead to nothing

An average property generates around 13 inquiries, leads to 12 viewings and ends with 1 serious buyer. For an office with 32 agents and 450 sales per year, that is 5,400 viewings, of which 4,500 viewings without a closing per year (the buyer views twice on average), or 11,250 working hours.

The solution is not fewer viewings but the right ones first. Whoever prioritizes the serious buyers sells faster and wastes less time on "property tourists".

4,500 viewings without a closing = €337,500 of lost working time

At 2.5h per viewing × €30 full cost/hour. This time is missing for acquisition and closings.

Inquiry data
→
Behavioral signals
→
Random forest
→
Lead score
→
Prioritization

02 Data basis: 6,000 inquiries, 12 months

Every inquiry with 14 behavioral signals, from first contact to closing

▸ Output
Leads: 6,000 | Buyers: 450 | Rate: 7.5%
6,000
Inquiries / year
7.5%
Conversion rate
92.5%
Non-buyers
450
Real buyers

03 Signals: what separates buyers from tourists

5 behavioral patterns nobody tracks systematically

Conversion rate by financing status
↳ The strongest signal

Leads with confirmed financing have an 18% conversion rate, 6× higher than leads without any financing information. But only 15% of inquiries contain this info. The model learns to compensate for the missing information from proxy signals: message length, follow-up-question frequency, and response time correlate strongly with financing readiness.

Conversion rate by response time (minutes)
↳ The speed indicator

Leads who reply within 30 minutes buy with 13% probability. After 4 hours the rate drops to 4%. Response time is no accident: it measures urgency and emotional commitment. A lead who reacts immediately has already decided to search.

Conversion rate: follow-up questions × budget match

04 Model: random forest with probability output

Not yes/no, but a purchase probability from 0-100%

26%
Hit rate in the top segment
Top 20%
contain 68% of buyers
0.79
AUC-ROC
↳ The practical translation

Hit rate in the top segment 26%: roughly one in four leads in the top segment buys, more than three times the base rate of 7.5%. The top 20% of leads contain 68% of all closings. That means if your team works the top-20% leads first, you reach two thirds of all buyers with a fifth of the effort. The remaining 80% of inquiries can be served with standardized replies and self-service.

Feature importance: what gives a buyer away

05 Lead dashboard: what it looks like day-to-day

Every inquiry automatically receives a score from 0-100

55
Hot leads (>70)
165
Warm leads (30-70)
280
Cold leads (<30)
500
Leads this month
L-04281 · Family · 3-room flat, Südvorstadt
Score: 87 / 100
Signals: financing confirmed · 3 follow-up questions in 2 days · response time 12 min · budget match 98% · owns a property (sale planned)
Recommendation: offer a viewing immediately. Priority 1.
L-04293 · Couple · house, southern suburbs
Score: 74 / 100
Signals: financing in progress · message of 120 words (detailed questions) · 5 listing views · weekend inquiry
Recommendation: viewing within 48h. Ask about financing status.
L-04310 · Single · penthouse, city center
Score: 45 / 100
Signals: no financing information · 8 inquiries on different properties · short standard message · but: budget match 102%
Recommendation: qualification call. Could be an investor, or a tourist.
L-04322 · Unknown · flat, Westend
Score: 12 / 100
Signals: 14 inquiries in 3 months (never a viewing) · no financing hint · response time >24h · budget match 68%
Recommendation: auto-responder. Do not prioritize.

06 Business impact: winning time back

Fewer viewings, more closings, higher agent satisfaction

Additional revenue from lead prioritization
€206,000
Additional revenue / year
-33%
Viewings without a closing (assumption)
CategoryAmount/yearMechanism
Additional closings€93,50011 saved buyers × €8,500 (speed to lead: reaction in minutes instead of hours; published: 21× higher qualification chance when answering within 5 minutes)
Viewing time saved€112,5001,500 avoidable viewings × 2.5 h × €30 (3,750 hours saved)
↳ The hidden effect

The biggest lever is not the time saved itself but the fact that hot leads get served faster. Currently a top lead waits just as long as a tourist, because inquiries are worked strictly in order of arrival. With scoring, the best buyer is contacted within minutes instead of after hours. That produces +11 closings per year, half of the buyers currently lost to faster competitors, at an average commission value of €8,500 per closing.

07 Next steps for your brokerage

From score to automatic prioritization

① CRM integration

An automatic score for every new lead in the CRM. Color coding: red/yellow/green. Sorting by score instead of arrival time.

② Auto workflows

Cold leads: automatic listing dispatch. Warm leads: qualification-call reminder. Hot leads: instant notification to the responsible agent.

③ Feedback loop

Every closing and every rejection flows back into the model. The score gets more accurate every month, a self-learning system.

All 6 modules: AI for real estate agents