Transcript Making the world a better place with SPSS: analysing & predicting

……. with SPSS
John Sauvé-Rodd
Datapreneurs™
How is the world made ‘better’ by SPSS?
 We can make sense of donor behaviour
 We recognise the best (and worst) donors
 We make budgets go further
 We make more money (net)
 We create INSIGHT
ASSESS November 2007
2
Agenda
1.
2.
3.
4.
5.
6.
7.
Charity fundraising in the UK
Analytics
Tools & skills
Common tasks
Advanced uses of SPSS
Q&A
Some light reading
ASSESS November 2007
3
But first - Let’s talk about ME!
• Veteran fundraiser and dataholic (with no
recovery plan)
• 25 years in the biz
• International fundraising consultant (no, really)
• Adore, love, addicted to SPSS
• Founder & Chair of the INSIGHT in Fundraising
Special Interest Group
• Want the truth?
– www.datapreneurs.net
ASSESS November 2007
4
Fundraising…..
It’s big business
190,000 registered charities
£40 billion (not a misprint)
raised annually
‘Top 20’ dominate revenue –
such as NSPCC, CRUK,
Salvation Army, Guide Dogs,
Save the Children
Known as the ‘third sector’ of
British civil society
More:
(http://www.charitycommission.gov.uk)
ASSESS November 2007
5
But where are the
TOOLS ?
Fundraising Analytics Evolution
•
•
•
•
•
STARTED: 1970s - direct mail marketing / mass markets / huge volumes / data
mining
GREW: phone, face to face, legacies, community fundraising
NEW: web fundraising
EMERGING: major gifts / the super-rich
YET TO COME: insight / melding of qualitative with quantitative research
ASSESS November 2007
6
Tools & Skills
6 x 9 = 42
CONSTRAINTS
• Fundraising databases can’t do analysis well
• EXCEL can’t handle large amounts of data
• raw SQL / VB programming v. tedious
• data transformation essential
• …. as well as statistical functionality
• budget/TCO always an issue
 SPSS is ideal (learning curve excepted)
ASSESS November 2007
7
Common tasks
Data validation & reformatting
Variable Information
Variable
id_no
gift_type
giftdate
gift
action_id
pay_mode
gift_type
po
monthly
giftyear
Position
1
2
3
4
5
6
7
8
9
10
Label
<none>
<none>
<none>
<none>
<none>
<none>
<none>
<none>
<none>
<none>
Measurem
ent Level
Scale
Scale
Scale
Scale
Scale
Nominal
Nominal
Nominal
Scale
Scale
ASSESS November 2007
Column
Width
8
8
10
8
8
9
7
7
10
10
Alignment
Right
Right
Right
Right
Right
Left
Left
Left
Right
Right
Print
Format
Write
Format
F7
F7
EDATE10
CCB5.2
F4
A1
A2
A1
F8
F8
F7
F7
EDATE10
CCB5.2
F4
A1
A2
A1
F8
F8
8
Data / file transformation
donor id no
1000029
1000029
1000031
1000031
1000031
1000031
1000031
1000031
1000031
1000031
1000031
1000032
1000032
gift date
27.11.1986
04.05.1983
29.06.1998
04.01.1994
27.11.1992
04.11.1991
18.12.1990
29.12.1987
09.07.1987
03.12.1986
04.05.1983
24.11.1992
15.10.1992
gift
€ 12.39
€ 4.83
€ 10.00
€ 10.00
€ 10.00
€ 10.00
€ 10.00
€ 10.00
€ 10.00
€ 10.00
€ 10.00
€ 12.39
€ 24.79
apeal code
3
1
2114
463
369
98
134
7
6
3
1
370
440
type
B
B
B
B
B
B
B
B
B
B
B
B
B
type
HZ
HZ
HZ
HZ
HZ
HZ
HZ
HZ
HZ
HZ
HZ
HZ
NH
PO
N
N
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
N
gift year
1986
1983
1998
1994
1992
1991
1990
1987
1987
1986
1983
1992
1992
monthly/non
Non monthly
Non monthly
Monthly
Monthly
Monthly
Monthly
Monthly
Monthly
Monthly
Monthly
Monthly
Non monthly
Non monthly
Let’s have a
look at some
syntax….
ASSESS November 2007
9
Advanced uses of SPSS in fundraising
• Complex data
• KPIs
– Multiple file joins for
advanced prospect
research
– ‘Stickiness’ analysis
• Predictive modelling
– Regression
– CHAID
–
–
–
–
–
–
–
ASSESS November 2007
Retention rates
Reactivation rates
File growth projections
Profitability*
Lifetime value
Donor life-cycles
Tenure
10
How is the world made ‘better’ by SPSS?
• We can make sense of donor behaviour:
– In a vast, swirling, ever-changing marketplace
• We generate genuine insight
• We recognise the best (and worst) donors
– And can thus meet their needs
• We make budgets go further
– Charities LOVE to save money
• We make more money (net)
– And this helps our cause and mission
ASSESS November 2007
11
Summary (applies to SPSS Base)
•
•
•
•
If we didn’t have SPSS …we’d have to invent it
SPSS’ flexibility is one of its strongest assets
TCO is good but the learning curve is steep
SPSS’ own training courses are poor:
– Because they are generic & not fundraising–focused
• Modular SPSS add-ons make analysis development
attainable
• Most of the ‘real’ stats applications in SPSS are unused
(and this is likely to continue for as long as fundraisers
remain a mathematically challenged group)
Q&A???
ASSESS November 2007
12
Some light reading
My 2007 research paper on Donor-level
Profitability (using SPSS & published by the
Institute of Direct Marketing) will be on the
ASSESS website for anyone foolhardy enough to
want to know more (20 pages / 6,000 words and
a lot of charts)
http://www.spssusers.co.uk/Events/2007/confprog.html
ASSESS November 2007
13
Words to
live by
ASSESS November 2007
14