Transcript Improving Time to Revenue

Improving Time to Revenue
Marcia Kadanoff
Firewhite Consulting Inc.
CEO & President
Intro
Why You’re Here
 More and more software is like milk
– Limited amount of time to get in, capture
revenue, maximize profits
6-6-1 economics
Open source
On demand
2
Intro
Why I’m Here
 Work at the intersection
– Direct marketing
– Marketing science
 CMO Magazine calls
– “New breed” of statistical marketers
3
Intro
Extreme Competition
Supply exceeds demand
– Capital is cheap
– Skilled labor is ubiquitous
– Infrastructure prices are dropping
– Excess supply of almost everything …
except customers
Source: “Extreme Competition”, McKinsey & Co. (Jan. 2005)
4
Intro
Traditional View of Marketing
Hierarchy of Effects
Purchase
Trial
Consideration
Interest
Awareness
5
Intro
Model
 Doesn’t fit
 Never validated in 30+ years
 Not for lack of trying
Qualitative
Focus Groups
Cognitive Scientists
• Eye Tracking
• Brain Scans
Brand Equity
Quantitative
Survey Research
Gallup Polls
Internet Panels
Marketing Scientists
• Choice Modeling
• MVT Testing
• Marketing-Mix Modeling
6
Intro
Extreme Makeover
“spark”
Customer
 Software
– Technology change
– Change in life stage
or at the company
– Regulatory change
– Reco of an expert
7
Customer Analytics
Build a Fact base
 Leverages behavioral data
– Accumulated in CRM and operational
systems at your company
• dB Analytics
• Statistics
– A/B and MVT testing
8
Customer Analytics
“Must Do” Analyses
Sources
Uses
Profiling
Relevant messages, offers,
media placement
Segmentation
Targeting
Product pricing, bundling
decisions
Customer Value*
How much can I afford to spend
to acquire and serve different
types of customers?
Choice Modeling*
Feature sets that go into a
product bundle
Pricing that optimizes profits
Market-Mix Optimization
Mix of spending that optimizes
revenue (e.g.)
*Not discussed today in the interest of time
See Appendix for add’l info
9
Best Practice #1
Profiling
 Take your customer file and match it
up against outside data sources
Source: Claritas (2005).
Also consider: Great Data, Dun & Bradstreet
10
Best Practice #2
Segmentation
 Groups “like” customers together
Customer Value
N
O
P
VP
MVP
Segment 1
Segment 2
Avoid
Acquire
Retain
Migrate
Retain
Clone
Collaborate
Segment 3
11
Best Practice #3
Time-based Analysis
 Profiling + Segmentation
– Analysis over time periods (this year vs. last)
– Can lead to some “Eureka” moments
– Indexing - no. of customers and revenue in two
categories (new/existing)
•
•
•
•
0-3 month
0-12 month
0-24 month
0-36 month
Index >100 growth
Index <100 shrinkage
12
Best Practice #3
Example
 Growth in 0-3 new buyers meeting a
particular profile
– Female, multi ethnic
– Younger than normal
– Educated but not technical
– Urban, Suburban
13
Best Practice #3
“Eureka Moment”
14
Customer Analytics
In the Future
 Analysts are predicting
– One product, customized on the fly to
meet the dynamic needs of customers
• Emma …
• George …
• Etc.
15
Customer Analytics
Love to detail
 Rest of analytic solutions but can’t in
a single hour
– “Answers” are in your customer dB
– Those that aren’t
Disciplined testing
• Test vs. Control - A/B testing
• MVT testing
16
Search Plus
Step 2 - Extreme Makeover
–
–
–
–
Technology change
Life stage
Regulatory change
Reco of an expert
“spark”
Customer
Hunt & Gather
Search front and center
 Not just search
–
–
–
–
–
Organic search (SEO)
Paid search (SEM)
RSS
Banner ads
Interactive
 FAQ
– Traditional Advertising?
– Non starter - cost reasons
– $150K per Q in sustained
spending
17
Search Plus
Best Practices
 #3 Measure result
– Using ROI not click throughs or conversion
rates
 #4 Test everything from end-to-end
Search term Text Ad  Landing Page
– Ideally using an MVT testing service like
Offermatica to speed up the process
18
Search Plus
Best Practices (cont)
 #5 Control you affiliates
– To avoid bidding against yourself
 #6 Leverage your fact base
– Expand search terms
– Guide media placements
– Determine messaging and visuals
19
Search Plus
Best Practices (cont)
 #7 Make your offers strategic
– Ideally, they should add and not subtract from your value
proposition
$39.99
$59.99
20
MOTS
Customer Experience
“spark”
Customer
Hunt & Gather
Experiment
 Is Make or Break
– “Moments of Truth”
•
•
•
•
Download
Installation
First-support incident
Purchase
Commit
21
Best Practice #8
Download
 Best Practices
– #8 Make download a simple one-step
process
22
MOTS
Creating An Account
 Is a No No
Requesting too much information
Too soon in the relationship
- Identifying info
- Profiling info
- Business critical info
- Opt-in to follow on communications
Will depress response
23
MOTS
Best Practices
 #9 Profiling
– Don’t collect information you can get
through other means
 #10 Ask for one behavior at a time
– Discipline based on “MOTS”
– Download  download + opt-in
24
Best Practice #11 - #15
Installation
 Murphy’s law
– Anything that can go wrong will go wrong
– Used to be true with download
– Burden has shifted to installation
 Best practices
#11 Use a commercial install product
#12 Don’t cripple your product
#13 Plan on nagging your customer
#14 Make the install window long enough
#15 Measure results using match back
25
Best Practice #16 - #18
First Support Incident
 This is a “Moment of Truth”
– Customer judgment is harsh, immediate
– You can win (or lose) a customer for life here
 Best Practice
– #16 Give prospects access to your support
forums -or- if you are just getting started FAQs
– #17 Be clear about the preferred method of
contact
– #18 Meet or exceed stated turnaround times
26
MOTS
Small Size
 Can be a substantial asset
– If you come clean
– Authenticity is rare
• “We’re a small company and depend on our
users as the first line of support”
• “I was amazed to find that the company
turned around a patch within 24 hours of my
making the issue known to them”
27
MOTS
“White Space”
“spark”
Customer
Hunt & Gather
 Best Practice
– #19 Short easy-to-scan
communications
– #20 Don’t even think about
violating customer’s privacy
Experiment
Commit
28
MOTS
True Commitment
“Hierarchy of Trust”
75% Emotional
Shared Values
25% Rational
Empathy
Responsiveness
Reliability
Functionality/Competence
Source: Global Fund for the Future (2005)
Adapted from a White paper “The importance of being Ernest”
Potential strength of relationship
 Is based on trust
29
After the Sale
Step 3 - Extreme Makeover
After the Sale
“spark”
Customer
Hunt & Gather
Experiment
Commit
– Acquisition
– Retention
– Migration
 Tactics
– Upgrade Mailings
– Collaboration
– Brand Advocacy
30
After the Sale
Upgrade Mailings
 They’re B-A-C-K
– EM is easy to ignore
– EM + DM will lift results by 20-50%
– Don’t sell features, sell benefits
• Example Recent SPSS mailing
– For mailings over 200K pieces consider
leveraging predictive analytics
• Takes into account the marketing mix and upgrades
you would have gotten anyway
• Example in the Appendix - courtesy of Quadstone
31
After the Sale
Collaboration
 Reduce cost to serve
 Up commitment
– Expert status - earned over time - protects
customer base from cherry picking
– Products “wrapped” more tightly around needs
– Facilitate brand advocacy/buzz/WOM
marketing
 Examples
– Threaded discussion board - pMachine
– Embrace blogosphere - MindJet
– Polling - Forrester
*See Appendix for some software solutions
32
After the Sale
Brand Advocacy
 Spread positive WOM on your behalf
Source: “The Marketing Value of Customer Advocacy”,
Wragg and Lowenstein, Ad Map, January 2005
33
After the Sale
Brand Advocacy
 Tightly related to customer value
Source: “The Marketing Value of Customer Advocacy”,
Wragg and Lowenstein, Ad Map, January 2005
34
Accountability
Step 4 - Extreme Makeover
Accountability
 Not a slam dunk
– Best customers - purchase through multiple
channels
– Tracking URLS - low incidence
– Source codes - only work 20-40% of the time
– Self-reported data on attribution - not
accurate
35
Accountability
Market-Mix Optimization
 Uses statistics to determine the best way to
allocate marketing dollars
– By product line
– By geography
– By media type
 “Secret weapon” used by companies
spending at least $5M on marketing
– Statistical methods of attribution
36
Accountability
Major ISV
 “Katmandu”
– Sells licensed client/server software into
the Enterprise
– Sells to line-of-business manager as well
as to IT
– Katmandu spends in excess of $50M per
year in the US on advertising and sales
promotion
37
Accountability
Inputs
 2 years of data
– Spending, sales, by product line by DMA
– Set of constraints determined by
management team
– Current allocation of marketing budget
38
Accountability
Media Plan Varied by DMA
Radio
TV
DM
Banner
EM
Search
Outdoor
9 0 %
8 0 %
7 0 %
6 0 %
5 0 %
4 0 %
Da y to n
3 0 %
C h ic a g o
2 0 %
A t la n t a
S e a t t le
1 0 %
D a lla s
0 %
R a d io
S a n D ie g o
T V
Ne w
Yo rk
D ir e c t M a il
De n v e r
Ba n n e r
B illin g s
E m a il
B o is e
Se a rc h
O u td o o r
M a d is o n
39
Accountability
Key Findings
 Reco Re-Allocation of Budget
% Budget Allocation
60%
50%
Before
Reallocated
40%
30%
20%
10%
0%
Radio TV DM EM Outdoor
40
Accountability
Business Impact
 ROI ~500% or 5x
30%
+23%
25%
+25%
20%
% Change
15%
10%
+6%
5%
0%
-5%
Conversion Units
Sold
Rate
-10%
-15%
-20%
-25%
Revenue
-19%
Cost to Acquire
41
Accountability
For Us Plebs
 With less than $5M to spend
– Match back - the Gold Standard
– Take list of people who downloaded trial
software and then match back to all purchasers
from all sources after allowing for a time lag
 Issues
– Appropriate time lag
– Fuzzy logic, phonetics
– Different match keys across different data
sources and/or accounting for missing data
42
Wrap Up
Extreme Competition
 Promised you 4 “big ideas”
– Marketing needs an extreme makeover
– In our extreme makeover we’d
• Focus on Customer Analytics not Market Research
• Put Search and Internet marketing front and center
• Optimize Customer Experience using what we know
works to drive downloads, installations, and purchase
• Leverage brand advocacy as a multiplier - but know
that we won’t get there without trust
• Invest in accountability, recognizing that marketing is
a process like any other business process
43
Wrap Up
Contact Info
 Marcia Kadanoff
 650.270.4309
 [email protected]
44
Wrap Up
This Presentation
– Contains copyrighted material and original
intellectual property produced by Marcia
Kadanoff on behalf of Firewhite Consulting, Inc.
– Feel free to use material here so long as you
attribute the source to:
• Marcia Kadanoff of Firewhite at www.firewhite.com
–  2005, all rights reserved.
45
Customer Analytics
Appendix
 Return on Investment
ROI = Units Sold * ASP * Margin - N(Cost to Acquire + Cost Serve)
Marketing Cost
– Where ASP is the average selling price of the products
and N is the number of customers
– Typically ROI is calculated on a segment-by-segment
basis
46
Customer Analytics
Appendix
 Customer Lifetime Value
CLV = m * r
1+i-r
–Where m = margin or profit from a customer per period (e.g. per year)
–Where r = retention rate, for example .8 or 80%
–Where i = discount rate, for example .12 or 12%
Source: Managing Customers as Investments, Gupta and Lehmann (2005)
47
Customer Analytics
Appendix
 To learn more
– About choice modeling
– See these pages on the Firewhite site
•
•
•
http://www.firewhite.com/services/npd_profit_maximizer
http://www.firewhite.com/thoughtleadership/index/choice_modeling/
http://www.firewhite.com/clients/cases/case_demand_planning
48
Software Tools
Appendix
 For Collaboration
– “First Look” of BrightIdea service
• Innovation Tools Weblog
http://www.innovationtools.com/Resources/ideamgmtdetails.asp?a=190
– CMO Magazine
• Requires free site registration
• Roundup of Idea Management tool
http://www.cmomagazine.com/read/090105/idea_sam
pler.html
– Informative
http://www.informative.com
49
After the Sale
Appendix
 This case courtesy of Quadstone
– Adapted by Firewhite
 Predictive Analytics
– Preparing for a major support mailing …
to get people to re-up through the mail
– Mail a random 50% of 1,000,000
customers
50
Case
Predictive Analytics
Sign up rate per month
5%
4%
3%
2%
1%
0%
Before
After
51
Case
Predictive Analytics
 Lots of divergent views
– The Database Marketing Manager says the
mailing worked
– The Director of Advertising says that it wasn’t
the mailing at all, but that it was the result of TV
advertising
– Customer support points out that the need for a
security audit (s.t. that’s free for customers that
re-up) was merchandised in an email
newsletter that went out to all customers
52
Predictive Analytics Case
Control Group
Sign Up Rate Per Month
5%
4%
3%
No mail group
2%
Mail group
1%
0%
Before
After
53
Predictive Analytics Case
Next Steps
 Great!
– Thanks to a “no mail” control group we
know mailing worked
– The Database Marketing Manager now
wishes to use predictive analysis to
improve the targeting of the next mailing
– He builds a decision tree . . .
54
Predictive Analytics Case
Decision Tree
Objective: Respond
Traditional CHAID analysis
5% of 1,000,000 mailed
49,873
Sex
Female
<40
4.1%
12,353
Male
4.3%
25,100
5.7%
24,773
Age
Age
>40
4.6%
12,747
<40
6.2%
12,321
Best mailing target
Men <40
Maximizes response
>40
5.2%
12,452
55
Predictive Analytics Case
Mailing ROI
Take-up Rate
Take-up Rate
No mail group
Mail group
Age
Age
Sex
18 - 39
40 - 65
Female
0.8%
0.4%
Male
2.8%
3.3%
Sex
18 - 39
40 - 65
Female
4.1%
4.6%
Male
6.2%
5.2%
Difference
18 - 39
40 - 65
Female
+3.3%
+4.2%
Male
+3.4%
+1.9%
56
Predictive Analytics Case
Problem
 As is often the case
– Decision tree identified lots of people
who signed up well after the mailing
– Raises questions about attribution of
results and who to target
57
Predictive Analytics Case
Solution
 Predictive Analytics
– Use results of controlled test to build a
predictive model, one that isolates the
impact of mail on uptake of support
renewals
58
Case
Predictive Analytics
Difference
18 - 39
40 - 65
Female
+3.3%
+4.2%
Male
+3.4%
+1.9%
• Objective: maximize response
• Predicts lift from mailing given particular
marketing mix
+3.2%
Sex
Female
<40
+3.3%
Male
+3.8%
+2.6%
Age
Age
>40
<40
+4.2%
+3.4%
Best mailing target
40+ women
Maximizes lift
>40
+1.9%
59
Case
Predictive Analytics
 Two different answers
– Who should we target?
• CHAID - men <40
• Predictive analytics - women 40+
– Answer is … it depends
• On whether the value of these customers is
equivalent or not
60
After the Sale
Appendix
 WOM marketing
– Available on B|NET
– Membership is free
– After you join, search for:
61