segmentation session 2014 washington nonprofit conferenc edited 2-19-14 cbk
DESCRIPTION
Direct marketing segmentation: format and packaging ideas as presented at 2014 DMANF Washington Nonprofit Conference.TRANSCRIPT
How Should We Segment Our Files in 2014 and Beyond?Cheryl Keedy, Senior Strategist, Production SolutionsJeff Regen, Vice President of Development, WETAKelly Townsend, Online Fundraising Manager, Humane Society of the United States
Agenda
Background: Why should we care so much about segmentation anyway?
Case studies from Humane Society of the United States
What do segmentation strategies mean for creative?
Ok, so how do we make advanced segmentation a reality?
Wrap-up
Background: Why should we care…
US population continues to grow…but giving not so much
Source: US Immigration Policy – Environmental Impact StatementNote: The Total Giving information was provided by AHP for Giving USA 2012Source: Giving USA - The Annual Report on Philanthropy 2012
4
Direct marketing metrics look bad
5
Source: Target Analytics donorCentrics
New donor trends look worse
6
Source: Target Analytics donorCentrics
Less Income x Lower % to Charity = Less Charity Dollars
7Source: Advisor Perspectives, Inc., 2013Source: Giving USA Foundation – The Annual Report on Philanthropy 2012
Source: Merkle, Inc. from Bureau of Labor Statistics, Consumer Expenditure Survey data. Giving to religious and educational institutions excluded
A deeper look shows younger generations are not giving as in the past
1991 1996 2001 2006 2011$0
$10,000,000
$20,000,000
$30,000,000
$40,000,000
$50,000,000
$60,000,000
$70,000,000
$80,000,000
Total Giving By Age Bracket
Under 25 25-34 Years 35-44 Years 45-54 Years 55-64 Years 65-74 Yerars 75 Years and Older
54%
35%45%
Collective revenue from < 55
11
Source: Merkle, Inc. from Bureau of Labor Statistics, Consumer Expenditure Survey data. Giving to religious and educational institutions excluded
Donors are different people who want different things – and communicate through different channels…
11
• Support charity they believe in• Support cause they believe in• Get benefits, premiums, gifts (and even mail)• Situational: Feel and see their tangible impact in a specific area(s)• Have different interests, preferences• Etc.
Source: Merkle, Inc. from Bureau of Labor Statistics, Consumer Expenditure Survey data. Giving to religious and educational institutions excluded
..and we are slowly depleting our files by optimizing to traditional retirees
11
• Support charity they believe in• Support cause they believe in• Get benefits, premiums, gifts (and even mail)• Situational: Feel and see their tangible impact in a specific area(s)• Have different interests, preferences• Etc.
Case studies from Humane Society of the United States
But first: What is The HSUS’ segmentation strategy?
• Segmentation approaches – aside from RFM – can be:– Issue-based (primarily online)– Geographic-based (online/offline)– Online behavior-based – e.g., actions taken; emails
clicked on– Based on offline purchasing history or donations to
other organizations (info from external vendors)• A word on the development of our two mail
programs, “Classic Marketing” and “Mixed Marketing”
$-
$0.20
$0.40
$0.60
$0.80
$1.00
$1.20
$1.40
Northern States Control Fleece Gloves Northern States Test Garden GlovesSouthern States Control Fleece Glove Southern States Test Garden Gloves
Net
/Pie
ce
11/12 HF Glove Series: Fleece vs. GardenNorth vs. South
Combined Results: Announce, Glove, Follow-Up
Testing premiums based on geographyClassic Marketing test (Fall 2012): Net/piece increased dramatically when the type of glove received matched the weather of the geographic region.
$(12)
$(10)
$(8)
$(6)
$(4)
$(2)
$-
ControlSouthern StatesNorthern States
Cost
/Don
or
5.02%
4.21%
5.34%
Fall 2012 Acq Glove Series: Fleece Gloves Control (Nationwide) vs. South vs. NorthResponse Rate Indicated
Testing premiums based on geographyClassic Marketing test (Fall 2012), cont.: Response rates & cost/donor results for the same test.
Testing premiums based on ageClassic Marketing test (Spring 2012): Cost/donor increased dramatically for the younger age groups who received jumbo address labels compared to older groups and the control.
2/2012 Acquisition Jumbo Labels – Age TestingResponse Percentage Indicated
$(40.00)
$(35.00)
$(30.00)
$(25.00)
$(20.00)
$(15.00)
$(10.00)
$(5.00)
$- Label Control (across the board)
Jumbo Labels (18-34 Years Old)
Jumbo Labels (35-54 Years Old)
Jumbo Labels (55+ Years Old)
Jumbo Labels to Expires (18-34 Years Old)
Jumbo Labels to Expires (35-54 Years Old)
Jumbo Labels to Expires (55+ Years Old)
Cost
/Don
or
0.92%
1.22%
1.92%
1.49%
1.56%1.90%
Adding online advocacy behavior to RFM
control x-section of file
0-12M w/Online Gift in 2012
12M w/Online Action in 2012
13M+ w/Online Action in 2012 $-
$5.00
$10.00
$15.00
$20.00
$25.00
$30.00
$35.00
$40.00
$45.00
$50.00
$0.27 $0.45 $0.36 $0.20
Average Gift
Net/Piece
Ave
rage
Gift
Classic Marketing test (2012): Avg gift & net/piece increased if an active or expired offline donor had also given online (of course) – and even if they had just taken an action (but not given) onlinein the same year as this mailing. Note the very high avg gift for expired offline donor action-takers.
Mixed Marketing test (Nov 2013): Ricky Bobby (year-end)-themed mailing sent to newly acquired online non-donor, 2+ action-takers at the beginning of the online year-end campaign in November 2013. Names had been acquired b/t May and October 2013.
Results:Qty
Mailed Qty Rtn % Rtn Avg Gift Net/M Net/Donor
Online Non-donor 2+ Action-takers Total
29,261 103 0.35% $32 ($153) ($43)
Multichannel 2013 donors ($10-499)
19,775 237 1.20% $63 $491 $41
Online 2013 donors ($10-499)
8,273 57 0.69% $80 $288 $42
Mail 2012-2013 donors ($10-499)
33,562 894 2.66% $38 $740 $28
Active sustainers 27,665 197 0.71% $63 $181 $25
Lapsed/Partially Lapsed Sustainers
1,292 5 0.39% $69 $3 $1
TOTALS 119,828 1,493 1.00% $58 $258 $16
Adding online advocacy behavior to RFM
One Time Pledges Sustainer Pledges
SEGMENT# 1x
Pledges%
PledgesAverage 1x Gift
# Sustaining Pledges
% Pledges
Average Sustaining
Gift
2013 Online FAP Donors 6 2.55% $45.83 20 8.51% $12.95
2+ Online FAP Action-Takers 37 2.54% $27.16 55 3.78% $9.58
Online FAP Email Clickers 16 2.73% $32.50 31 5.28% $12.52
TOTALS 59 2.61% $35.16 106 5.86% $11.68
What? Telemarketing for sustaining giftsWhen? After running a two-month multichannel farm animal-themed campaign – with the primary channels being online (email, web, advertising, social media, mobile), but including a mailingWho? (1) Donors who had given by mail or online, (2) 2+ farm animal action-takers, & (3) online campaign email (non-donor) “clickers”
Adding online advocacy & email behavior to RFM for telemarketing
Calling campaign results (Oct-Nov 2013):
What do segmentation strategies mean for creative and results? A broader industry scan
Rails-to-Trails Conservancy MKDM
• Test #1: 2013 RTC Summer Appeal – Retaining state participation in the Recreational Trails Program Tested an abundance of state specific personalization State nameMiles of trails in stateLong Description of trailName of well known trailGovernor’s name*Mailed July/August with no back end premium offer
RTC Summer Appeal Letter/Reply 14 personalized state specific touch points
Rails to Trails – Summer Appeal2013 State Versioning Results
Summer Appeal
Reponse Rate Avg. Gift Income/M Net/M CTRAD
2013 Appeal (Personalized, state-specific campaign) 3.75% $47.96 $1,800.72 $1,430.53 $0.21
2012 Actual 2.53% $54.91 $1,388.19 $1,079 $0.22
2011 Actual 2.50% $43.85 $1,097.18 $656 $0.40
2010 Actual 3.06% $45.19 $1,384.10 $1,050 $0.33
Notes:
1) Significantly higher response rate in 2013 compared to all previous years. Huge increase in income/M and net/M as well.
2) 2013 Appeal raised as much total as 2012, while reducing mail volume by 25%.
Rails-to-Trails ConservancyMKDM
• Test #2: 2013 RTC April Appeal – tested region specific guidebooks on outer envelope, reply form and URL depending on the state the donor lived in vs. a generic guidebook version
April Guidebook
Regional Testing Response Avg. Gift Cost/M* Income/M Net/M CTRAD
Midwest 3.41% $49.49 $468.89 $1,688.69 $1,219.80 $0.28
Midwest Generic 3.51% $50.69 $468.97 $1,778.05 $1,309.08 $0.26
Northeast 3.66% $45.87 $465.53 $1,681.29 $1,215.76 $0.28
Northeast Generic 3.47% $45.40 $465.60 $1,575.44 $1,109.84 $0.30
New England * 4.18% $49.94 $470.09 $2,089.51 $1,619.42 $0.22
New England Generic * 3.41% $49.18 $470.09 $1,679.02 $1,208.93 $0.28
No Region 2.40% $47.34 $427.65 $1,137.47 $709.82 $0.38
Subtotal: Region-Specific 3.68% $48.11 $467.76 $1,771.57 $1,303.82 $0.26
Subtotal: Generic 3.47% $48.19 $467.82 $1,672.91 $1,205.10 $0.28
* Only test cell with a statistically significant lift in response rate at a 95% confidence level.
Rails-to-Trails April AppealResults: Only one test cell with a significant lift – New England
AARP Foundation
• Test: 2012 Annual Fund: Segmented donors who responded to a previous appeal effort on the same topic and incorporated test language referencing same.
“as someone who supported our Annual Fund last year….”
AARP Foundation Test Results
Annual Fund (April 2012)
• Test - 9.77% response rate; $17.02 Av Gift• Control - 8.92% response rate; $16.90 Av Gift• Results – Test language = +10% in response and +1% Av
Gift
In 2013 AARP Foundation incorporated the strategy into 3 appeals:• Annual Fund (April); Calendar (September); and Year-End
Challenge (December)
The Kennedy Center - Avalon
• Goal: Improve performance of in house acquisition lists
• Test: Use ‘performance teasers’ tailored to past ticket purchases. Three most popular categories: Dance, National Symphony Orchestra, and Theater.
• Incentive: Members-only ticket priority (ability to purchase tickets in advance of the general public).
• Results: An average 40% lift in response
Control Teaser
Performance Tailored Teaser Copy NSO, Dance & Theatre
Schultz & Williams
• Test #1 - Children’s Hospital – color/gender• Test #2 - The Cornell Lab of Ornithology –
Chronic Non-Responders• Additional Beyond RFM Test Findings
Children’s Hospital: Selecting females and giving them a very feminine package—monarch, light pink, hallmark-style labels increased response over the full campaign by 51%.
Gender
Conservation Organization:Identified chronic non-responders to annual “pillar” campaign. Tested those names on a more standard appeal. More standard appeal generated 2X response and added another 15% of revenue to the appeal timeslot.
Chronic Non-Responder
Schultz & Williams: Other Test Findings
• National conservation organization: added a buckslip to donors who lived near the programmatic work being highlighted. Lifted response by 17%.
• Regional zoo: special membership offer to those prospects living within the same county. Lifted response by 42%.
• National social justice organization: created an online welcome series that asked for top interest out of three choices. Next message led with that interest. Open rates of over 50%.
• National social justice organization: added a buckslip about the IRA tax-deduction to donors who qualified (age = 70.5 and higher). Overall IRA revenue was up significantly.
• National advocacy organization: tested an increased font size (14-point) for all constituents. Long-time members increased their response by 18%; no discernable impact on newer, younger members.
Loyal Donor Segmentation - RenewalDefenders of Wildlife
‘you would continue to stand with us as you have for so many years’
‘You have stoodwith us for so long and with such incredible commitment and dedication.’
Loyal Donor Control Loyal Donor Test
Health Organization Test Findings – NNEHouse File
• Donors notifying the organization of a connection to the disease:• Have a 57% higher response rate.• Have a 39% increase in average gift.
• Donors with an email address on file have:• A 15% lower response rate in the mail.• Average gift is 55% greater in the mail.• Are worth 32% more than donors without email address.
• Males give at the same response rate as females, but average gift is 20% higher.
NNE strategies undertaken based on these trends:
• Deeper lapsed audiences are selected if they have an affinity to the disease or an email address on file.
• Creative doesn’t ask the donor how the disease affects them if we already know. We instead use the space to state other facts about the organization.
• Copy treatment within the letter recognizes the relationship the donor has with the organization.
• Different gift asks are tested and implemented for males versus females.
• Other gender based creative variation has yielded results:• Use of graphics and types of graphics (floral vs. nostalgic)• Letter length (shorter length = men)• Gift asks (higher for men)
Ok, so people are different. But how can we segment and treat them differently?
Segment and personalize based on data – 3 types of variables1. What they do (with your nonprofit) – Giving history
2. What they do (with your nonprofit) – All other interactions1. Email activity (opens, clicks, actions)2. Social media3. Website and other online activity4. Attend events5. Call 800# or email Member Services6. Interests7. Etc.
3. Who they are (demographics, psychographics, 3rd-party data)
Creating dynamic segmentation and personalization
Giving history
• Giving level• Frequency• New, first year, second year, multi-year, many year, lapsed, non-donor
Relationship to
org
• Donor, activist, volunteer, event participant, buyer, etc.
Channel
• Mail, telemarketing, email, website, social media, face-to-face, etc.Motivation
• Motivation: Philanthropic, transactional, etc.• Issue interest area: Dogs vs. cats, public radio vs. TV, environmental issue area, etc.D
emographics
• Traditional retirees vs. affluent professionals, etc. Gender, geography, etc.
Selecting the right segmentation strategy for your organization – Key questions1. What sets of variables are the primary drivers of performance for your org? E.g., for
WETA: • Giving history / potential• Demographic-psychographic segments• Public television vs. public radio• Public television interest area – News/public affairs, drama, kids programs, etc.• Motivation: Premium-driven vs. subscription-driven vs. philanthropic
2. What is independent impact on performance of each set of variables?
3. What does coverage look like for each set of variables (broader is better – e.g., RFM is typically best place to start)
4. How difficult will it be to implement adjustments to messaging, cadence, etc.? E.g., for WETA, television vs. radio and motivation both have good coverage and impact performance
5. What’s the right staging / progression? Note: Starting off with all at once is typically too much. 1-2 sets of variables should be dominant form of segmentation
Building out dynamic segment and individual-level strategies1. Build out segment-level strategies for major segments (for both acquisition and
existing file):1. Messaging2. Offer3. Channels4. Cadence
2. Modeling and advanced analytics can inform selects – and consider model variations by segment
3. Invest in each individual / segment based on potential value
4. Use data to drive personalization at individual level (e.g., based on person’s behavior)
• Cheryl Keedy• Senior Strategist Direct Response &
Marketing Formats• Production Solutions• [email protected]• 703-402-7477
• Jeff Regen• Vice President, Development• WETA• [email protected]• 703-998-2653
• Kelly Townsend• Online Fundraising Manager• Humane Society of the United
States• [email protected]• 240-620-4990