View
267
Download
0
Embed Size (px)
DESCRIPTION
Citation preview
Customer Intelligence
Ludo LonginAdministrateur Délégué Direct Social Communications
My favourite song: Queen
Geert VerstraetenPartner at Python Predictions
My favourite song: Absynthe Minded
Ludo LonginDirect Social Communications
How knowing donors helps in growing donors
Geert VerstraetenPython Predictions
Analytics in Fundraising
= communication agency
°1985
1 activity:fundraising for humanitarian organisations
Team:12 enthousiastic people
Direct Social Communications nv
Health FoodBelgium Belgian Assocation for Burn Injuries
Belgian Cystic Fibrosis AssocationKom op Tegen Kanker (Cancer Assoc.)
Food BanksRestaurants of the Heart
World Mercy ShipsChain of HopeDamien FoundationMedics Without Vacation
Children AnimalsBelgium Collective Research and Expression
Youth VillagePelicano Foundation
World If Child Help Veterinarians Without Borders
Handicap PeopleBelgium Blind Care Light & Love Flemish Autism Association
World Handicap InternationalSensorial Handicap Cooperation
Pilots Without BordersMamas for AfricaThe Voice of the Lebanese WomenFriends of Sister Emmanuelle
± 400 fundraising campaigns using dm
> 8,000,000 letters to private individuals
> 1,500,000 inserts in newspapers
> 600,000 donations per year
> 24,000,000 euros
Female
60+ years old
religious
Fully owned appartement/house
Adult children have left the house
Typical dm-donor in Belgium
Response rate in housemailing:◦ 8 %
◦ 10 %
◦ 12 %
◦ 14 %
◦ More than 15 %
Normal results of dm-campaigns
Response rate in Acquisition campaigns:◦ 2 %
◦ 2,5 %
◦ 3 %
◦ 4 or More %
Normal results of dm-campaigns
Recruitment campaignsWhere do we get our donors from?
D.S.C.
± 220,000addresses
CONSUDATA+ criteria
6 millionaddresses
Otherdatabases
Our Prospection ApproachCustomer/Donor Cloning
Use of analytics: advanced (predictive models, similarity models)
Core business: Predictive AnalyticsSince 2006Based in BrusselsReferences:
Python Predictions
STEP Select the best donors:
Cloning in a Nutshell
1
STEP Calculate area conversion rate:
Cloning in a Nutshell
2
STEP Calculate area profile
Cloning in a Nutshell
3Residents
Housing
Neighbourhood
STEP Build Predictive Model
Cloning in a Nutshell
4Residents
Housing
Neighbourhood
0,0%
1,0%
2,0%
3,0%
4,0%
0 50000 100000 150000 200000 250000 300000
CloningGeo SegmentationAge
Number of Households Targeted
Response
STEP Validate Predictions
Cloning in a Nutshell
5
0%1%2%3%4%5%6%7%8%9%
0‐4 5‐9 10‐14 15‐19 20‐24 25‐29 30‐34 35‐39 40‐44 45‐49 50‐54 55‐59 60‐64 65‐69 70‐74 75‐79 80‐84 86‐89 90‐94 >95
Age categoryHouseholds we would not target
Households we would target
STEP Validate Predictions
Cloning in a Nutshell
5
8%26%
0 1 2 or more
Number of bathrooms
9%
21%
N/A < 35 m2 35 till54 m2
55 till84 m2
85 till104 m2
105 till124 m2
> 125m2
House surface
STEP Validate Predictions
Cloning in a Nutshell
5 Households we would not target
Households we would target
1,98% 2,29%2,71%
Age > 55y CloningTop 100.000
CloningTop 10.000
Response Percentage
21,6 €25,5 € 28,6 €
Age > 55y CloningTop 100.000
CloningTop 10.000
Average Donation Amount
improvement of 37% in response rate
improvement of 32% in donation amount
Campaign Results (First Test)
0,43 €0,58 €
0,78 €
Age > 55y CloningTop 100.000
CloningTop 10.000
Break(revenue per letter sent)
improvement of 82% in revenue per letter sent
Current and Future Usage
Campaign Results (First Test)
Geert [email protected]
Tel +32 2 762 69 00
Python Predictionswww.pythonpredictions.com
QUESTIONS? Or later?
Ludo [email protected] +32 2 280 00 74
Direct Social Communicationswww.dsc.be