Marcia Kadanoff Firewhite Consulting Inc. CEO & President Improving Time to Revenue.

Marcia Kadanoff

Firewhite Consulting Inc.

CEO & President

Improving Time to RevenueImproving Time to Revenue

2

Why You’re HereWhy You’re Here

More and more software is like milk– Limited amount of time to get in, capture

revenue, maximize profits

6-6-1 economicsOpen sourceOn demand

Intro

3

Why I’m HereWhy I’m Here

Work at the intersection– Direct marketing– Marketing science

CMO Magazine calls– “New breed” of statistical marketers

Intro

4

Extreme CompetitionExtreme Competition

Supply exceeds demand– Capital is cheap– Skilled labor is ubiquitous– Infrastructure prices are dropping– Excess supply of almost everything …

except customers

Intro

Source: “Extreme Competition”, McKinsey & Co. (Jan. 2005)

5

Traditional View of MarketingTraditional View of Marketing

Awareness

Interest

Consideration

Trial

Purchase

Hierarchy of Effects

Intro

6

Doesn’t fit Never validated in 30+ years Not for lack of trying

Intro

QuantitativeSurvey ResearchGallup PollsInternet PanelsMarketing Scientists

• Choice Modeling• MVT Testing• Marketing-Mix Modeling

QualitativeFocus GroupsCognitive Scientists

• Eye Tracking• Brain Scans

Brand Equity

ModelModel

7

Extreme MakeoverExtreme Makeover

Software– Technology change– Change in life stage

or at the company– Regulatory change– Reco of an expert

Customer

“spark”

Intro

8

Build a Fact baseBuild a Fact base

Leverages behavioral data– Accumulated in CRM and operational

systems at your company

• dB Analytics• Statistics

– A/B and MVT testing

Customer Analytics

9

““Must Do” AnalysesMust Do” Analyses

Sources UsesProfiling 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 timeSee Appendix for add’l info

Customer Analytics

10

ProfilingProfiling

Take your customer file and match it up against outside data sources

Best Practice #1

Source: Claritas (2005).

Also consider: Great Data, Dun & Bradstreet

11

SegmentationSegmentation

Groups “like” customers together

N O P VP

Segment 1

Segment 2

MVP

Segment 3

AvoidAcquireRetainMigrate

RetainClone

Collaborate

Customer Value

Best Practice #2

12

Time-based AnalysisTime-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 growthIndex <100 shrinkage

Best Practice #3

13

ExampleExample

Growth in 0-3 new buyers meeting a particular profile

– Female, multi ethnic– Younger than normal– Educated but not technical– Urban, Suburban

Best Practice #3

14

““Eureka Moment”Eureka Moment”Best Practice #3

15

In the FutureIn the Future

Analysts are predicting– One product, customized on the fly to

meet the dynamic needs of customers• Emma …• George …• Etc.

Customer Analytics

16

Love to detailLove 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

Customer Analytics

17

Step 2 - Extreme Makeover Step 2 - Extreme Makeover

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

– Technology change– Life stage– Regulatory change– Reco of an expert

Customer

Hunt & Gather

“spark”

Search Plus

18

Best PracticesBest 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

Search Plus

19

Best Practices (cont)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

Search Plus

20

Best Practices (cont)Best Practices (cont)

#7 Make your offers strategic– Ideally, they should add and not subtract from your value

proposition

$39.99 $59.99

Search Plus

21

Customer ExperienceCustomer Experience

Is Make or Break– “Moments of Truth”

• Download• Installation• First-support incident• Purchase

Customer

Hunt & Gather

“spark”

Experiment

Commit

MOTS

22

Best Practice #8

DownloadDownload

Best Practices– #8 Make download a simple one-step

process

23

Creating An AccountCreating An Account

Is a No No

Requesting too much informationToo soon in the relationship

- Identifying info- Profiling info- Business critical info- Opt-in to follow on communications

Will depress response

MOTS

24

Best PracticesBest 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

MOTS

25

InstallationInstallation

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

Best Practice #11 - #15

26

First Support IncidentFirst 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

Best Practice #16 - #18

27

Small SizeSmall 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”

MOTS

28

““White Space”White Space”

Best Practice– #19 Short easy-to-scan

communications– #20 Don’t even think about

violating customer’s privacy

Customer

Hunt & Gather

“spark”

Experiment

Commit

MOTS

29

True CommitmentTrue Commitment

Is based on trust

Functionality/Competence

Reliability

Shared Values

Responsiveness

Empathy

Pot

entia

l str

engt

h of

re

latio

nshi

p“Hierarchy of Trust”

Source: Global Fund for the Future (2005)Adapted from a White paper “The importance of being Ernest”

75% Emotional

25% Rational

MOTS

30

Customer

Hunt & Gather

“spark”

Experiment

Commit

Step 3 - Extreme MakeoverStep 3 - Extreme MakeoverAfter the Sale

After the Sale– Acquisition– Retention– Migration

Tactics– Upgrade Mailings– Collaboration – Brand Advocacy

31

Upgrade MailingsUpgrade 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

After the Sale

32

CollaborationCollaboration

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

After the Sale

33

Brand AdvocacyBrand Advocacy

Spread positive WOM on your behalf

Source: “The Marketing Value of Customer Advocacy”, Wragg and Lowenstein, Ad Map, January 2005

After the Sale

34

Brand AdvocacyBrand Advocacy

Tightly related to customer value

Source: “The Marketing Value of Customer Advocacy”, Wragg and Lowenstein, Ad Map, January 2005

After the Sale

35

Step 4 - Extreme MakeoverStep 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

Accountability

36

Market-Mix OptimizationMarket-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

Accountability

37

Major ISVMajor 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

Accountability

38

InputsInputs

2 years of data– Spending, sales, by product line by DMA– Set of constraints determined by

management team– Current allocation of marketing budget

Accountability

39

Media Plan Varied by DMAMedia Plan Varied by DMA

M a d is o n

B o is e

B illin g s

D e n v e r

N e w Y o r k

S a n D ie g o

D a lla s

S e a t t le

A t la n t a

C h ic a g o

D a y t o n

R a d i o

T V

D ir e c t M a i l

B a n n e r

E m a i l

S e a r c h

O u td o o r

0 %

1 0 %

2 0 %

3 0 %

4 0 %

5 0 %

6 0 %

7 0 %

8 0 %

9 0 %

RadioTVDMBannerEMSearchOutdoor

Accountability

40

Key FindingsKey Findings

Reco Re-Allocation of Budget

0%

10%

20%

30%

40%

50%

60%

Radio TV DM EM Outdoor

Before

Reallocated

% B

ud

get

All

oca

tio

n

Accountability

41

Business ImpactBusiness Impact

ROI ~500% or 5x

-19%

+6%

+23% +25%

-25%

-20%

-15%

-10%

-5%

0%

5%

10%

15%

20%

25%

30%

% C

ha

ng

e

Cost to Acquire

ConversionRate

UnitsSold

Revenue

Accountability

42

For Us PlebsFor 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

Accountability

43

Wrap Up

Extreme CompetitionExtreme 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

44

Contact InfoContact Info

Marcia Kadanoff 650.270.4309 [email protected]

Wrap Up

45

This PresentationThis 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.

Wrap Up

46

AppendixAppendix

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

Customer Analytics

47

AppendixAppendix

Customer Analytics

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%

Customer Lifetime Value

Source: Managing Customers as Investments, Gupta and Lehmann (2005)

48

AppendixAppendix

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

Customer Analytics

49

AppendixAppendix

For Collaboration– “First Look” of BrightIdea service

• Innovation Tools Webloghttp://www.innovationtools.com/Resources/ideamgmt-details.asp?a=190

– CMO Magazine • Requires free site registration • Roundup of Idea Management tool

http://www.cmomagazine.com/read/090105/idea_sampler.html

– Informativehttp://www.informative.com

Software Tools

50

AppendixAppendix

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

After the Sale

51

Predictive AnalyticsPredictive Analytics

0%

1%

2%

3%

4%

5%

Sign up rate per month

Case

Before After

52

Predictive AnalyticsPredictive 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

Case

53

Control GroupControl Group

0%

1%

2%

3%

4%

5%

No mail group

Mail group

Before After

Sign Up Rate Per Month

Predictive Analytics Case

54

Next StepsNext 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 . . .

Predictive Analytics Case

55

Decision TreeDecision Tree

49,873

25,100

12,353 12,747

24,773

12,321 12,452

Objective: RespondTraditional CHAID analysis5% of 1,000,000 mailed

SexFemale Male

Age<40 >40

Age<40 >40

4.3% 5.7%

4.1% 4.6% 6.2% 5.2%

Age<40 >40

Predictive Analytics Case

Best mailing targetMen <40Maximizes response

56

Mailing ROIMailing ROI

Take-up Rate

No mail group

 Age

Sex 18 - 39 40 - 65

Female 0.8% 0.4%

Male 2.8% 3.3%

Take-up Rate

Mail group

 Age

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%

Predictive Analytics Case

57

Predictive Analytics Case

ProblemProblem

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

58

SolutionSolution

Predictive Analytics– Use results of controlled test to build a

predictive model, one that isolates the impact of mail on uptake of support renewals

Predictive Analytics Case

59

Predictive AnalyticsPredictive Analytics

+3.2%

+3.8%

+3.3% +4.2%

+2.6%

+3.4% +1.9%

• Objective: maximize response• Predicts lift from mailing given particular

marketing mix

SexFemale Male

Age<40 >40

Age<40 >40

Difference

18 - 39 40 - 65

Female +3.3% +4.2%

Male +3.4% +1.9%

Case

Best mailing target40+ womenMaximizes lift

60

Predictive AnalyticsPredictive Analytics

Case

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

61

AppendixAppendix

WOM marketing– Available on B|NET– Membership is free– After you join, search for:

After the Sale