Barak Libai lecture WIM UK April 2010

Prof. Barak LibaiProf. Barak Libai 1

ASSESSING THE VALUE OFCUSTOMERS’ WORD–OF–MOUTH

Prof. Barak Libai

Tel Aviv University

WOM UK , London, April 2010

By 2010, most mangers are informed about the power ofcustomer word of mouth

Prof. Barak Libai

We also know that assessing the economiccontribution of word of mouth is critical

Understanding the real valueof customers

What is the “social value” of aperson?

Prof. Barak Libai 4

value of advertisingrealUnderstanding the

Prof. Barak Libai 5

• Valuing investments in social networksand social media in general

Planning and valuing WOMcampaigns E.g., should we target “influentials”?

Prof. Barak Libai

Yet for many, the way WOM turns into ROIstays a mystery

Prof. Barak Libai

Here are I will argue that:

The issue is not trivial. No “one number” or “one equation” youneed to know

We need to “speak CRM”. The WOM value measure (“socialvalue”) should be monetary: The effect on the value of the cashflows from other customers

We need to see the larger network. Because word of mouthcreates a complex effect, its impact should take into account thelarger social system, and not only close neighbors

We want to understand the value of time. Early is often MUCH better.

The value a customer brings us via WOM

Current Approaches

Indirect measures do no assess monetary value

Direct aggregate measures

Measures for a “social value” of a customer

Prof. Barak Libai 8

Indirect measures

A) How do customers talk? Or are willing to recommend?

E.g, Net Promoter Scores

Prof. Barak Libai 9

Prof. Barak Libai 1010

Keiningham, B Cooil, TW Andreassen, L Aksoy - Journal ofMarketing, 2007

Conclusions are not straightforward

B) How do customers listen? How did you hear of us? How much were you affected by word of

mouth?

Possibly used with brand equity measures

Recent research found that the long term value ofcustomers that had arrived via WOM was higher thanthat of customers that had arrived via advertising.

Prof. Barak Libai 11

Direct “aggregate” measuresHow word of mouth affects overall sales

Experiments Before and after a WOM campaigns

In different areas

Prof. Barak Libai 12

Advanced statistical regressions

Especially as data on sales, WOM activities and othercustomer data is available from websites

Issues of “identification”: For example, WOM may be affected by advertising which may

affect the level of WOM

People may behave the same because there are similar, or wereaffected by the same external phenomenon, and not becausethey talked

Prof. Barak Libai 13

The problems with the previousapproaches

We still do not understand the “how” in terms of valuecreation

Monetary value of the individual is not assessed

Prof. Barak Libai 14

Prof. Barak Libai

The “social value” of a customer

The extra monetary value a customer adds (or subtracts)to the firm due to social interactions with others

Three issues to be discussed: Lifetime Value, Network,Time

Issue 1: “where the money is”?

Prof. Barak Libai 16

The “regular” profitability of customers

Customer Lifetime Value (CLV) - the net present value(NPV) of the future cash streams from a customer

Prof. Barak Libai 17

Time

$

$

$

$

$

The money that comes from a customer’sword of mouth is in other customers’

regular lifetime value!

Prof. Barak Libai 18

$

$

$

$

$

$

$

$

$

$$

$

$

$

$

$

$

$

Often, we consider the social effect of in acquiring newcustomers What is their lifetime value?

Do friends have a similar lifetime value?

But it can also be an issue of costs E.g., via reduced customer support in online communities

A case of customer “development”- cross sell or up-sell

Or affecting the probability of customer defection The effect of retention (defection) on customer lifetime value is

very strong!

Prof. Barak Libai 19

Where’s the money?

Reichheld: The Loyalty Effect

Recent research in the telecommunication industry

ברק ליבאי'פרופ-שיווק בתקשורת חברתית 21

Exposure to Defecting Neighbors - Defecting and Non-DefectingCustomers

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

Mar08 Apr08 May08 Jun08 Jul08 Aug08 Sep08 Oct08 Nov08 Dec08

Average number ofdefecting neighbors

Month

Defectors

Non-Defectors

Prof. Barak Libai 22

How the social effect of a defection decreases with time

It can also be mere acceleration of cashflows

“Discount rate” plays a large role in customer lifetimevalue A customer starting to buy today may be worth much more than

a customer starting to buy later

Prof. Barak Libai 23

$

$

$

$

$

$

$

$

$

$

My conclusion on this issue

The same way that CRM people start “talking social”

Prof. Barak Libai 24

Social people should start “talking CRM”

Issue 2: Where is the network?

Prof. Barak Libai 25

Prof. Barak Libai

The complexity of calculating person A’s social value

A

B C

Is it the “lifetime value” of the people s/he affected?

D

Prof. Barak Libai

The complexity of calculating person A’s social value

A

B C

How many of them could we get via advertising?

If this is the case, we should look only at savings in advertisingexpenses

D

Prof. Barak Libai

The complexity of calculating person A’s social value

A

B C

Person A ’s word of mouth may create a “chain effect”beyond the neighbors

D

Prof. Barak Libai

The complexity of calculating person A’s social value

A

B C

If A would not talk with B, C may do it in a later time

Is social value about “customer acquisition” or“customer acceleration”?

D

Prof. Barak Libai

The complexity of calculating person A’s social value

A

B C

Can we add the social value of A and B ?

D

Prof. Barak Libai

Conceptually, the real social value of a customerif weonlyor a group of customers) can be assessed(

Let the customer disappear And measure the change of the net present value of the whole social

system

Prof. Barak Libai

Can it really be calculated?

We are working on it !

I’ll next present a possible approach.

Prof. Barak Libai

Stage A- Collect data on real social networks

Prof. Barak Libai

Stage b: Based on these networks create simplesimulations in which products are sold to

connected customers

For example: what would happen if a new product wouldbegin to grow on such networks

Individual level simulations in which a “would-be-world”is created are sometimes called agent based models

Prof. Barak Libai

Stage 3: Conduct experiments

What is the profitability (NPV) of the system

If person A is there, or is not there

If we target “influentials” or random customers

If competition is strong or not

In the absence of tools such as agentbased models

Try to better understand social network analysis

Prof. Barak Libai 36ברק ליבאי'פרופ-שיווק בתקשורת חברתית 4

Prof. Barak Libai 37

•Degree Centrality -the number of direct connections a node has

•A node with high degree centrality is a “Hub”

•We can also differentiate between in and out degree

•Eigenvector centrality – Gives weights to the centrality of the nodes thatare direct connection ( the “degree”)

•Google’s “PageRank” is a variation on this

•Closeness centrality – the sum of shortest paths to all others

•The shorter the better

•Betweenness Centrality - How many shortest paths between otherspass through that person

Who is important to us?

Issue 3: The Value of Time ?

Prof. Barak Libai 38

My (and others’) research has repeatedly indicated thatbeing early in the market has long lasting effects due toword of mouth

Prof. Barak Libai 39

Prof. Barak Libai 40

0

100

200

300

400

500

600

700

0 1 2 3 4 5 6 7 8 9 10

Loss($)

Year Since Introduction

Indirect effect

direct effect

VALUE OF ONE LOST CUSTOMER IN THE ONLINE BANKING INDUSTRY

WOM

Regularpurchases

Prof. Barak Libai

Some recent results on the social value of WOM“seeding” programs

Using agent based models based on 12 real networks

WOM programs create a real pioneering advantageamong competitors. The social value of a program canbe five time as much when a competitor does not have aprogram

It is very worthwhile to be first, and alone!

Thank you !

Questions? libai@post.tau.ac.il

Prof. Barak Libai 42

How much value a customer creates viaword of mouth?

What is “value”?

Even before that: What is “word of mouth”?

Prof. Barak Libai 43

The classic view of “word of mouth”

Prof. Barak Libai 44

?Online or Offline

The vast majority of recent knowledge on socialinteractions comes from online environments

Yet much of the action may still be offline

Prof. Barak Libai 45

Other, 2%

E-mail, 3%

IM/Text, 3%Chat/Blog, 1%

Face-to-Face,

73%

Phone, 17%

Source: OMD/Keller Fay Group proprietary report based on TalkTrack®, June 5th 2006 through February 3, 2008

Source: TheKeller Fay group

“Organic” or “Amplified”?

Can we transferknowledge from firmincentivizedcampaigns to “natural”word of mouth ?

Prof. Barak Libai 46

Kumar, Petersen and Leone, HBR2007

What about “observational learning?”

We may be seriouslyunder-estimating thevalue of socialinteractions!

Prof. Barak Libai 47

I use the term “word of mouth” (WOM)

But the issues covered largely include various kinds ofsocial interactions

Prof. Barak Libai 48