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 - the buying decision starts with the photo
Yet no agent measures which photos actually work
The buyer's first decision - "Do I click on this listing?" - is made in seconds. Cover photo, price, location. After that, photo quality decides whether an enquiry comes in or not. Still, listing photos are produced by most agents out of habit rather than for effect.
The data for optimisation already exists: portal statistics (views, click rate, enquiry rate), photo characteristics (brightness, perspective, room type), text length, keyword density. We connect them into a model that says: "This listing performs 34% below its potential - here is why."
Top-quartile listings: more than twice as many enquiries as the bottom quartile (4.8 vs. 2.1 per 1,000 views)
For an identical property, identical price, identical location. The only difference: the listing design.
Portal statistics
→
Photo features (CV)
→
Text features (NLP)
→
Performance score
→
Optimisation advice
02 Analysis - what top listings do differently
2,000 listings analysed quantitatively (3 years, incl. unsold properties) - 12 measurable quality factors
import numpy as np, pandas as pd
from sklearn.ensemble import GradientBoostingRegressor
# Features per listing
features = [
# Photo features (computer vision)
'anzahl_fotos', # How many images?
'foto_helligkeit_avg', # Average brightness (0-255)
'foto_schaerfe_score', # Laplace variance as sharpness measure
'hat_drohnenfotos', # Aerial shots present?
'hat_grundriss', # Floor plan in the listing?
'titelbild_typ', # Exterior/interior/drone/garden
'fotos_pro_raum', # Coverage: how many rooms shown?
# Text features (NLP)
'text_laenge_woerter', # Description length
'emotional_score', # Emotional vs. factual language
'keyword_count', # SEO-relevant keywords
'hat_energieausweis_info',# Mandatory disclosures complete?
# Structure features
'preis_sichtbar', # Price shown or "on request"?
]
# Target: enquiries per 1,000 views (normalised by location/price)
# R² = 0.38 - listing quality explains 38% of enquiry variance
# after controlling for location, price and property type (model assumption)
print("Top drivers of enquiry rate:")
print("1. Cover photo type (exterior in daylight: markedly more enquiries)")
print("2. Number of photos (sweet spot: 18-22)")
print("3. Floor plan present (+28%)")
print("4. Photo brightness (>160/255 optimal)")
print("5. Text length (sweet spot: 180-280 words)")
Enquiry rate by listing attribute - top vs. bottom
↳ The cover photo decides
Listings with an exterior shot in daylight as the cover photo attract markedly more enquiries than those with interior shots - eye-tracking studies confirm the cover-photo mechanism; the size of the effect is an assumption of the model calculation. But: for penthouse apartments the roof-terrace perspective works better. For detached houses the garden view dominates. There is no "best photo" - there is the best photo for this property type.
Enquiry rate vs. number of photos - the optimum
↳ More is not always better
The enquiry rate rises up to 18-22 photos, then falls again. The reason: too many photos signal "the agent is trying too hard" or overwhelm the buyer. Fewer but better photos is the formula. And: a missing floor plan costs 28% of enquiries (published up to +93%, Rightmove - we deliberately set our figure below that); the simplest quick win in the entire listing.
03 Business impact - better listings, more enquiries, faster sales
Every additional qualified enquiry accelerates the sale
Additional revenue from listing optimisation
€76,500
Additional revenue / year
+9
Closings saved / year (assumption)
| Category | Amount/year | Mechanism |
|---|
| Closings saved | €76,500 | 9 saved closings × €8,500 - better listings save weaker properties from the mandate expiring (assumption; mechanism published: professional photos sell 32% faster, Redfin/VHT) |
↳ The quick win
The three immediately actionable measures: (1) A floor plan in every listing (+28% enquiries, cost: €0). (2) Cover photo rule: always an exterior shot in daylight as the default. (3) At least 18 photos, maximum 22. These three rules alone raise the enquiry rate by an estimated 20% - without a single euro of extra cost.