magnifying glasses and crystal balls
DESCRIPTION
Use your data to raise more today and predict the future (Presented at BBCON 2013)TRANSCRIPT
04/08/2023 #bbcon 1@BI_JCA ~ [email protected]
Magnifying Glassesand Crystal Balls
Use your data to raise more today and predict the future
04/08/2023 #bbcon 2@BI_JCA ~ [email protected]
Bob Dillane, Director, Enterprise Information Systems
Lebanon Valley College
Cari Maslow, Senior Director, Donor Relations and Membership
Carnegie Museums of Pittsburgh
Teri Morrow, Membership Director
Appalachian Mountain Club
Stephanie Reyes, Manager, Business Intelligence Group
JCA
Steve Beshuk, Director, Business Intelligence Group
JCA
YOUR PANEL
04/08/2023 #bbcon 3@BI_JCA ~ [email protected]
• What is Business Intelligence?
• Case Studies
- Lebanon Valley College: The Power of a Data Warehouse
- Appalachian Mountain Club: Tracking Members Source Using BI
- Carnegie Museums of Pittsburgh: The Power of Cubes
• Predictive Analytics
• Q&A
SESSION AGENDA AND GOALS
04/08/2023 #bbcon 4@BI_JCA ~ [email protected]
WHAT IS BUSINESS INTELLIGENCE?
04/08/2023 #bbcon 5@BI_JCA ~ [email protected]
Business intelligence is a set of theories, methodologies, processes, architectures, and
technologies that transform raw data into meaningful and useful information for
business purposes.
BI IS…
04/08/2023 #bbcon 6@BI_JCA ~ [email protected]
Business intelligence is a set of theories, methodologies, processes, architectures, and
technologies that transform raw data into meaningful and useful information for
business purposes.
BI IS…
04/08/2023 #bbcon 7@BI_JCA ~ [email protected]
NON-TECH BI ARCHITECTURE
Warehouse
Data InformationKnowledge
CRM
04/08/2023 #bbcon 8@BI_JCA ~ [email protected]
DATA WAREHOUSE
Sales
PK id_product
id_client
id_client group
id_shop
id_supplier
id_date
id_country
id_suppliername
id_supplieraddress
id_shop
ETL
CRM
04/08/2023 #bbcon 9@BI_JCA ~ [email protected]
ANALYTIC CUBES
04/08/2023 #bbcon 10@BI_JCA ~ [email protected]
VISUALIZATION
04/08/2023 #bbcon 11@BI_JCA ~ [email protected]
TYPES OF VISUALIZATION
TABLES CHARTS & GRAPHS HEAT MAPS
TIME SERIESTREE GAUGES
04/08/2023 #bbcon 12@BI_JCA ~ [email protected]
• The Challenge- The data landscape has changed – there’s a lot more of it now
- Constant struggle to grow revenue and keep costs in check
- Today’s donors are demand results
- Unfortunately, many nonprofits are behind the curve
• The Opportunity- Today’s BI gives you the tools to leverage your untapped “data asset”
- Sophisticated BI does not have to mean expensive
- BI can pay for itself – results!
BI AND NON-PROFITS
04/08/2023 #bbcon 13@BI_JCA ~ [email protected]
CASE STUDY: THE POWER OF A DATA WAREHOUSE
04/08/2023 #bbcon 14@BI_JCA ~ [email protected]
LEBANON VALLEY COLLEGELebanon Valley College in Annville, Pa.,
welcomes 1,600 full-time undergraduates
studying more than 30 majors, as well as self-
designed majors. Founded in 1866, LVC has
graduate programs in physical therapy,
business, music education, and science
education. Annville is 15 minutes east of
Hershey and 35 minutes east of Harrisburg;
Philadelphia, Washington, D.C., and
Baltimore are within two hours.
FAST FACTS
• Guaranteed degree completion in 4 years
• 3 out of 4 LVC students choose to live in
excellent, safe, guaranteed campus housing
• Low student-to-faculty ratio of 13 students
to each professor allows for personal
interactions and a customizable education
• High-achieving students, 1/2 of whom were
in the top 20% of their high school class,
continually teach each other while learning
together
• 98 % of students receive some form of
financial assistance with an average aid
package of more than $24,500
04/08/2023 #bbcon 15@BI_JCA ~ [email protected]
• A transaction database is designed for efficient data entry
• A data warehouse is designed for ease and speed of reporting
• These are two very different design goals
TRANSACTION DATABASE VS. DATA WAREHOUSE
04/08/2023 #bbcon 16@BI_JCA ~ [email protected]
• Technical Concerns- Unless you have a large database, technology is not a big concern
• Practical Concerns- Can I get answers with a data warehouse that I can’t (efficiently) get without
one? (the answer is virtually always “yes”)
- Will those answers improve our ability to perform our mission?
DO YOU NEED A DATA WAREHOUSE?
04/08/2023 #bbcon 17@BI_JCA ~ [email protected]
• SQL code to query event participant attributes directly from the Raiser’s Edge database:
COMPLEXITY OF THE RAISER’S EDGE
04/08/2023 #bbcon 18@BI_JCA ~ [email protected]
• Same query against the warehouse:
SIMPLICITY OF A DATA WAREHOUSE
04/08/2023 #bbcon 19@BI_JCA ~ [email protected]
• Reports we can’t do, or can’t do easily in RE
• Reports that need data from outside RE
• Perform calculations and import data back to RE
• Create custom gift analysis warehouse generated from JCA Answers warehouse
GENERAL USES FOR OUR WAREHOUSE
04/08/2023 #bbcon 20@BI_JCA ~ [email protected]
• This is on fund categories, not funds. Our goals are by category instead of individual funds.
FUND REPORTING
04/08/2023 #bbcon 21@BI_JCA ~ [email protected]
DONOR GIVING SUMMARY
04/08/2023 #bbcon 22@BI_JCA ~ [email protected]
• The Gift Analysis table has information about gifts, not donors.
• ConstitGifts table has links donors and gifts, with a flag for the type of linkage (hard credit, soft credit, match credit, etc.)
• Using these tables, we can quickly answer questions such as:- “How much did each donor give to the Mund Buidling Fund, the Annual Fund
and any endowment funds (3 separate totals) between July 17th and September 12th?”
• For frequently asked amounts (annual fund, etc.), we generate totals nightly and import them to RE as constituent attributes. This extends the capability of query and export in RE.
USING GIFT ANALYSIS WAREHOUSE
04/08/2023 #bbcon 23@BI_JCA ~ [email protected]
• We also use the warehouse to calculate the alumni/parent year code that is used with constituent names.
• We analyze:- What degrees and graduation years the individual has- What undergraduate degrees and graduation years the person’s children
have- We calculate an alumni/parent code and import it back to RE as an Add/Sal
OTHER WAREHOUSE USE IN SUPPORT OF RE DATA
04/08/2023 #bbcon 24@BI_JCA ~ [email protected]
• SQL Server Reporting Services is our formatted report writer• Entrinsik Informer as our primary add hoc query tool• Crystal Reports for some formatted reports, but we are phasing this
out• Excel with Microsoft Query to access the warehouse• JCA Answers as our primary data warehouse- Warehouse rebuild runs at 8:00 p.m.- Various calculations are run and exported from the warehouse- Calculations are imported into The Raisers Edge- Warehouse is rebuilt again at 7:00 a.m.
REPORTING TOOLS USED AT LEBANON VALLEY COLLEGE
04/08/2023 #bbcon 25@BI_JCA ~ [email protected]
CASE STUDY FROM APPALACHIAN MOUNTAIN CLUB: TRACKING NEW MEMBERS BY SOURCE USING BUSINESS INTELLIGENCE
04/08/2023 #bbcon 26@BI_JCA ~ [email protected]
• Helps people protect, enjoy, and understand the mountains, forests, waters, and trails from Maine to Washington, DC.• 86,000 members; 16,000 volunteers; 20,000 advocates• 12 chapters• Backcountry huts, camps & campsites, and front country lodges• Maintain over 1,800 miles of trail; 350+ miles of the AT• 8,000 volunteer- and staff-led activities
ABOUT APPALACHIAN MOUNTAIN CLUB
04/08/2023 #bbcon 27@BI_JCA ~ [email protected]
• Implemented in 2010• Reporting- Appeals- Member number - Member retention- New members- To Finance
BUSINESS INTELLIGENCE AT AMC
Leaving Madison Spring Hut Photo by Chris Lawrie, AMC New Hampshire Chapter
04/08/2023 #bbcon 28@BI_JCA ~ [email protected]
THE “OLDEN DAYS”
SourceNew
M’brshps RevenueAvg. Dues Payment
Total Cost to Acquire
Cost/ Member
Net/ Member
Collective Buying 351 $10,118 $28.83 $15,515 $44.20 ($15.38)
Direct Mail 7,742 $250,936 $32.41 $619,509 $80.02 ($47.61)
Email/Web 4,641 $242,555 $52.26 $21,693 $4.67 $47.59
Gift Memberships 659 $35,048 $53.18 $5,374 $8.15 $45.03
Miscellaneous 988 $31146 $31.52 $27,897 $28.24 $3.29
Reservations 852 $52,277 $61.36 $10,651 $12.50 $48.86
Telemarketing to Formers 1,374 $65,841 $47.92 $66,906 $48.69 ($0.78)
Total 16,607 $687,921 $41.42 $767,544 $46.22 ($4.79)
• Standard 2010 new member report by source
04/08/2023 #bbcon 29@BI_JCA ~ [email protected]
• What is new member retention by source?• What is retention rate, added giving, cost to renew?• Incorporated BI data into long-term reporting• Appeals coded with type and channel (AQDM, RNWB)• Type and channel split in BI—track at the channel level
THE UNKNOWNS
04/08/2023 #bbcon 30@BI_JCA ~ [email protected]
2011 RAW BI DATA
Source1st Year
Retention Pool1st Year
Retained%
Retained1st Year
Retained $Avg. Dues Payment
Collective Buying 277 74 27% $2,150 $29.05
Direct Mail 7,742 3,367 47% $113,566 $31.23
Email 174 49 28% $2,331 $47.57
Miscellaneous 646 236 37% $9,851 $41.74
Reservations 979 324 33% $19,166 $59.15
TM to Formers 1,900 602 32% $28,865 $47.95
Web 4,459 1,828 41% $87,939 $48.11
Total 16,177 6,750 42% $263,866 $39.09
04/08/2023 #bbcon 31@BI_JCA ~ [email protected]
ADDING CONTEXT WITH BI
Source2011 #
RetainedM’brshp +
Annual Fund $2012 #
RetainedM’brshp +
Annual Fund $Avg. Dues Payment
Collective Buying 74 $2,328 33 $1,542 $37.24
Direct Mail 3,367 $135,693 2,443 $124,894 $39.82
Email/Web 1,877 $106,114 1,120 $79,875 $53.49
Gift Memberships 100 $5,254 85 $5,291 $50.51
Miscellaneous 236 $13,684 141 $11,541 $46.27
Reservations 324 $22,766 207 $15,981 $62.14
TM to Formers 602 $41,651 269 $22,496 $53.93
Total 6,850 $327,489 4,298 $261,620 $45.74
04/08/2023 #bbcon 32@BI_JCA ~ [email protected]
• Re-calculated retention as of 2012 and re-calculated ROI.
FILLING IN THE DETAILS
SourceYear 3
RetentionTotal Cost
M’brshp Revenue
Total Ann. Fund
GivingNet/New Member
Year 3 M’brshp
ROI
% Ann Fund
Giving
Collective Buying 9% $17,825 $13,497 $491 ($10.93) $0.78 4%
Direct Mail 32% $657,151 $461,779 $49,744 ($18.81) $0.78 11%
Email/Web 24% $46,681 $392,731 $35,813 $82.28 $9.18 9%
Gift Memberships 13% $8,225 $43,670 $1,923 $56.70 $5.54 4%
Miscellaneous 14% $33,212 $47,521 $8,850 $23.44 $1.70 19%
Reservations 24% $15,039 $84,305 $6,718 $89.18 $6.05 8%
Telemarketing to Formers 20% $75,938 $109,212 $20,776 $39.34 $1.71 19%
Total 26% $854,071 $1,152,714 $124,315 $25.47 $1.50 11%
04/08/2023 #bbcon 33@BI_JCA ~ [email protected]
• 2013- Dropped Groupon/Living Social - Reduced cost of direct mail acquisition- Hired staff to boost non-direct mail/non-telemarketing efforts
• Growth in web/email sales slow• New members from guest reservations up nearly 300%• Memberships purchased by visitors at AMC huts, camps, and lodges
fewer than 100 in 2010 now up to nearly 900 in 2013.
• Tracking 2011 and 2012 new members• Guest stay data not available• No volunteer data in database
IN THE MEANTIME…
04/08/2023 #bbcon 34@BI_JCA ~ [email protected]
• Cut direct mail acquisition by 20% • Use Google Grant (awarded 8/2013) • Develop plan to boost retention and Annual Fund support from gift
memberships• Assess non-Membership department sales for 2014 growth potential.
PLAN FOR 2014
Crisp ColorsPhotograph by Nicholas Gagnon, AMC New Hampshire Chapter
04/08/2023 #bbcon 35@BI_JCA ~ [email protected]
CASE STUDY: THE POWER OF CUBES
04/08/2023 #bbcon 36@BI_JCA ~ [email protected]
•Founded in 1895•4 Distinct Museums
– Carnegie Museum of Art– Carnegie Museum of Natural History – Carnegie Science Center– The Andy Warhol Museum
•Serve 1.3 million people annually
CARNEGIE MUSEUMS OF PITTSBURGH
04/08/2023 #bbcon 37@BI_JCA ~ [email protected]
• Pressure to increase revenue with shrinking resources• Mountains of data in Raiser’s Edge but no easy way to get answers to
questions• Answers we had weren’t actionable and instead just led to more
questions• No easy way to measure strategy performance which led to a
culture of adding initiatives but not discontinuing any already implemented
THE NEED FOR BUSINESS INTELLIGENCE
04/08/2023 #bbcon 38@BI_JCA ~ [email protected]
• Projecting revenue was the impetus for taking what seemed to be a simple question further
• With an on-going, multi-hit renewal series, we needed to estimate not only what percentage would renew but when they would renew
• Knew the makeup of the population would impact the results but had no way of knowing which characteristics were most impactful
THE QUESTION OF RETENTION RATE
04/08/2023 #bbcon 39@BI_JCA ~ [email protected]
The retention rate is 63%
SO:
# of members x average gift x 63% =
Overall Renewal Revenue Projection
BUT:• Makeup of population impact both
the monthly and the year-to-year results
• Needed to predict when in the solicitation cycle renewals would happen
WHY DOES IT MATTER?
04/08/2023 #bbcon 40@BI_JCA ~ [email protected]
Multi vs. First Year Members• Use different renewal strategies• Run monthly query off an attribute set
for segmentation of the first direct mail hit
• Didn’t answer the “when” question • Confirmed theory that the makeup of
the monthly populations mattered
BEFORE JCA ANSWERS: THE EVOLVING RETENTION SPREADSHEET
04/08/2023 #bbcon 41@BI_JCA ~ [email protected]
TimingNext we started running multiple queries and plugging the counts into this spreadsheet data table
BEFORE JCA ANSWERS: THE EVOLVING RETENTION SPREADSHEET
04/08/2023 #bbcon 42@BI_JCA ~ [email protected]
The retention rate at each time point then calculated into summary tables
BEFORE JCA ANSWERS: THE EVOLVING RETENTION SPREADSHEET
04/08/2023 #bbcon 43@BI_JCA ~ [email protected]
• It worked but…• Labor Intensive• Risk of human error high• Not an easily transferable task• Technical: had to make friends in IT
SO…
04/08/2023 #bbcon 44@BI_JCA ~ [email protected]
• No clarity around the month to month variations- Large Variances- Not Consistent Year over Year
• Multi vs. First was not enough population segmentation
• Still left us without answers when asked “Why…”
• AND then we changed our upgrade solicitation method which further skewed the model
BUT…
04/08/2023 #bbcon 45@BI_JCA ~ [email protected]
• How many times do I have to get a member in the door?• Does renewal rate change if I have an active email address?• Does it matter whether they visit one, two or all three museum sites?• Are the increasing number of member events making a difference?• Retention gains in one year don’t appear to hold, why not?
Everything we did to try to impact retention was still just an educated stab in the dark.
MORE QUESTIONS
04/08/2023 #bbcon 46@BI_JCA ~ [email protected]
WITH JCA ANSWERS WE COULD DATA MINE
What’s going on with the Individuals?
Does it matter how many
years?
Is there a difference
between joins & rejoins?
True of all multi years
or does upgrading impact?
Do the rates by category stay
consistent month to month?
Does the purchase
method impact the rate?
What about timing?
04/08/2023 #bbcon 47@BI_JCA ~ [email protected]
By month:
DELVING INTO THE MONTH OF EXPIRE
Membership category:
PREM19%
FMLY50%
DUAL23%
INDL3%
SENR5%
January
PREM22%
FMLY54%
DUAL17%
INDL2%
SENR4%
March
04/08/2023 #bbcon 48@BI_JCA ~ [email protected]
• Number of Gifts:
DELVING INTO THE MONTH OF EXPIRE
• Renewed vs. Upgraded:
• Greater Detail on Family Level
04/08/2023 #bbcon 49@BI_JCA ~ [email protected]
IDENTIFICATION OF SIGNIFICANT FACTOR
• Purchase Method:
DM - Direct Mail52%
January
DM - Direct Mail38%
March
04/08/2023 #bbcon 50@BI_JCA ~ [email protected]
• Ability to identify which factors determine the larger variances in retention
• Role out testing that focuses on chosen factors and monitor results- Years of Membership- Join vs. Rejoin- Direct Mail Retention by Zip Code
• Ability to target specific groups of people with different offers & communication- Non-visiting members
WITH JCA ANSWERS
04/08/2023 #bbcon 51@BI_JCA ~ [email protected]
• Nugget: Getting them into their 3rd year is key
• New Strategy: Extend 1st year discounting to 2nd year members
IS THERE A MAGIC NUMBER OF YEARS?
2012
Why is the pattern changing in 2012?
20112010
04/08/2023 #bbcon 52@BI_JCA ~ [email protected]
• Nugget: Depends on the level of membership
• Strategy Idea: Switch telemarketing acquisition of lapsed members from Family & Premium to Dual, Individual and Senior levels
JOIN VS. REJOIN
04/08/2023 #bbcon 53@BI_JCA ~ [email protected]
• Nugget: Zip codes with strong purchase data don’t always have strong retention rates
• New Strategy: Work with direct mail firm to use data from Answers when making list buys
ACQUISITION LIST PURCHASE
04/08/2023 #bbcon 54@BI_JCA ~ [email protected]
OTHER CASE STUDIES
04/08/2023 #bbcon 55@BI_JCA ~ [email protected]
• Implemented JCA Answers for The Raiser’s Edge and Gateway Galaxy about 4 years ago
• Have an integrated BI environment
• Incorporated budget information into data warehouse
• Have been able to better understand deferred revenue reporting
• One department alone saw savings of 500 hours in staff time a year due to efficiencies made possible by JCA Answers.
AQUARIUM
04/08/2023 #bbcon 56@BI_JCA ~ [email protected]
• Converted to The Raiser’s Edge about 4 years ago
• Came to JCA initially for a data warehouse
• Now have an integrated BI infrastructure that combines their Raiser’s Edge and Ticketmaster VISTA ticketing data
• Starting to do more work in Predictive Analytics – initial savings: $20,000 annually
ART MUSEUM
04/08/2023 #bbcon 57@BI_JCA ~ [email protected]
PREDICTIVE ANALYTICS
04/08/2023 #bbcon 58@BI_JCA ~ [email protected]
LOOKING INTO THE CRYSTAL BALL
04/08/2023 #bbcon 59@BI_JCA ~ [email protected]
PREDICTIVE ANALYTICS IS HERE
• Rise in Google Searches of “Predictive Analytics”
04/08/2023 #bbcon 60@BI_JCA ~ [email protected]
• Rise in Google Searches of “Predictive Analytics” versus “Business Intelligence”
PREDICTIVE ANALYTICS IS HERE
04/08/2023 #bbcon 61@BI_JCA ~ [email protected]
Predictive analytics is business intelligence technology that uses predictive models built from your
data to make predictions about the future.
WHAT IS PREDICTIVE ANALYTICS?
04/08/2023 #bbcon 62@BI_JCA ~ [email protected]
• Start with a business question
How do I raise more money?
versus
How can I increase the number of membership renewals?
WHERE DO WE BEGIN?
04/08/2023 #bbcon 63@BI_JCA ~ [email protected]
• Do we have all the data we need?• Is our data usable?
PREPARATION IS KEY
04/08/2023 #bbcon 64@BI_JCA ~ [email protected]
• Explore your data• Keep your business question in mind
LET’S EXPLORE
04/08/2023 #bbcon 65@BI_JCA ~ [email protected]
• Models use patterns found in your data to identify risks and opportunities.
• Models apply scores to your constituents, which can help guide your strategy for improving outcomes
PREDICTIVE MODELS
04/08/2023 #bbcon 66@BI_JCA ~ [email protected]
• Naïve Baise- Used for classification
PREDICTIVE MODELS
04/08/2023 #bbcon 67@BI_JCA ~ [email protected]
• Decision Trees- Used for classification, regression and association
PREDICTIVE MODELS
04/08/2023 #bbcon 68@BI_JCA ~ [email protected]
• Clustering- Used for segmentation
PREDICTIVE MODELS
04/08/2023 #bbcon 69@BI_JCA ~ [email protected]
• Models are a work in progress.
• Test before deploying your model.
IMPROVE YOUR MODEL
04/08/2023 #bbcon 70@BI_JCA ~ [email protected]
• Reliable – Your predictive model must be accurate.
• Repeatable – You need to be able to use your model more than once.
• Relatable – You need to understand the results.
THE THREE R’S OF PREDICTIVE ANALYSIS
04/08/2023 #bbcon 71@BI_JCA ~ [email protected]
QUESTIONS?
04/08/2023 #bbcon 72@BI_JCA ~ [email protected]
• Leave us your card• Send us an email at [email protected]• Stop by our booth: 107/109• Tweet us @BI_JCA• Go to our website: www.jca-answers.com
FOR A COPY OF THIS PRESENTATION