Consumer Querimetrix

Option 1

presents

New Rules, New Tools!

CONVENTION TODAY. INVENTION YESTERDAY

An illustration—the Beer Can

PULL-TAB

THE CONVENTION TODAY!

Today’s convention—the Pull Tab

- Before pull-tabs (1935-60)

- made of tin (not Aluminum)- Opened with a churchkey-Churchkey: forgotten tool

Pull-tab = Engineer + picnic + beer – churchkey- First invented in 1960

The Moral—not about beer cans

But the story: - not about beer cans

The main plot: -The obsolete tool

(churchkey)

TODAY’S CONVENTION

wasYESTERDAY’S INVENTION

New Rules, New Tools

WHEN RULES CHANGE,

TOOLS NEED TO CHANGE!

CHANGING GROUND RULES ARE CHANGING

The India story—a ‘digital’ consumer story

Urban user: 70% access internet dailyRural user: 60% use 2-3 times a week

Source: IAMAI, IMRB International, TRAI

Search queries—capture the Zeitgeist Moment of truth gives way to ‘micro-moments’

EXAMPLE: THE EVOLVING CAR-BUYING JOURNEY

Car-buying journey—Then and Now!ONLINE EVALUATION CHANGING THE RULES

The New Consumer Decision Journey

PURCHASING FUNNEL— a simple straight routeNEW CONSUMER DECISION JOURNEY—more complex and unpredictable

Purchasing funnel theory gives way to complexity

The Active Evaluation Phase

• Active evaluation adds 2-3 more brands to your initial consideration

• Only BUDGET and COLOR are non-negotiable

• At least ONE brand initially considered gets eliminated

• 54% of car buyers change their consideration set during their research phase

The car buyer is unpredictable and price-sensitive

Google searches—in evaluation phase

40%YoY growth in car queries

100+queries every second on Google

search

8-10 times dailyOn average every car buyer uses Google search during his car buying journey

50% car buyers research themselves

25-30%additional car buyers have their friends and family research for the car purchase

Data on consumer behavior—a problem of plenty

DECODING GOOGLE TRENDS: BIG DATA PROBLEM

Solving A Big Data Problem

• Are search queries related to real-world activity?• If so, which search categories/queries are relevant?

Deriving consumer insights from search queries

Consumer activity needs to be associated with search queries

In A Big Data WorldAnalyze Google Trends with traditional tools

QUANT

OLD TOOLS

BIG DATA

WEEK #1

Big Data—solution or problem?Traditional rule-based systems are not very effective

1) ADAPTABILITY

TRADITIONAL DECISION SUPPORT TOOLSSHORTCOMINGS:

GAME-CHANGER?INFORMATION OVERLOAD?

2) SCALE

QUERIMETRIX: BIG DATA MEETS MACHINE LEARNING

Consumer QuerimetrixConsumer insights using Google Trends

TELEMETRY (REMOTE MEASURE)

CONSUMER SEARCH QUERIES

Uses framework proposed by Hal Varian and Steven Scott‘Predicting the Present using Bayesian Structural Time Series’ whitepaper series (2009-14)

The need for Machine LearningThe Swiss Army Knife of Business Intelligence tools

ADAPTABLE: multi-utility

SCALABLE: ASIMO to Predictive text

Consumer QuerimetrixBig Data meets Machine Learning

Consumer QuerimetrixWhat is the value proposition?

• Identifies proxy indicators

• Delivers on-ground insights

• Nowcasts to stay ahead of the herd

• Catches inflection points earlier

• Derives business insights using digital footprints


• identifies proxy indicators





Querimetrix—as a proxy indicatorHow are the online used car classifieds shaping up?• OLX is a clear leader from Google Trends

dataOLX leads Quikr in auto-related queries on Google

Normalized query trends for India car classifieds, 2004-current

Source: Google Trends, Kotak Institutional Equities

3799

038

130

3827

038

410

3855

038

690

3883

038

970

3911

039

250

3939

039

530

3967

039

810

3995

040

090

4023

040

370

4051

040

650

4079

040

930

4107

041

210

4135

041

490

4163

041

770

4191

042

050

4219

042

330

0

20

40

60

80

100

120 carwale cartrade olx quikr







The Maruti Baleno car launchWhat does Google Trends tell us about car launches?

21,000+Bookings in 10 days after

launch

56,000footfalls in 10 days of launch

9,000November 2015 sales (>i20)

40,000+Bookings since launch

Querimetrix—an ear to the groundComparing recent car launches on Google Trends

4085

340

895

4093

740

979

4102

141

063

4110

541

147

4118

941

231

4127

341

315

4135

741

399

4144

141

483

4152

541

567

4160

941

651

4169

341

735

4177

741

819

4186

141

903

4194

541

987

4202

942

071

4211

342

155

4219

742

239

4228

142

323

0102030405060708090

100

Honda Amaze Ford Ecosport Hyundai Creta Maruti Suzuki S-CrossHyundai Xcent Renault Kwid Maruti Suzuki Baleno

Car model launches result in large online inquiry of launched model

Google Trends search volumes of recently launched cars in India, 2011-15


KwidSecond-best launch

Baleno

Best launch of the year

Querimetrix—on ground insightsDoes higher digital curiosity imply anything?

A linear relationship is evident between advance bookings and search volumes

The relationship between search queries and advance bookings during car launches


10 20 30 40 50 60 70 80 90 1000

5,000

10,000

15,000

20,000

25,000

30,000 Maruti S-Cross Hyundai Creta Hyundai Xcent Honda AmazeFord Ecosport Trend Linear (Trend) Maruti Baleno

Indexed Google query volumes

Adva

nce

book

ings higher

search-share

higher

bookings

Querimetrix—an ear to the groundIs it merely a reflection of traditional media spends?

Advertising spends may not always guarantee digital buzz

The relationship between traditional media spend and advance bookings during car launches

Source: Google Trends, TAM, Kotak Institutional Equities

10 20 30 40 50 60 70 80 90 1000

10,00020,00030,00040,00050,00060,00070,00080,00090,000

100,000

Maruti S-Cross Hyundai Creta Hyundai Xcent Honda AmazeFord Ecosport Maruti Suzuki Ertiga

Indexed Google search share

Adve

rtisin

g sp

ends

(Rs.

'000

)

Notes:

(a) Advertising spends are based on TAM estimates of TV and Print spends during launch week.

NOlinear relationship













Querimetrix—nowcastingThe BSTS algorithm catches inflection points earlier

Querimetrix nowcasts ~195,00 passenger car sales in January

Consumer Querimetrix nowcasts of total passenger car sales, 2009-16 (units)

Source: SIAM, Kotak Institutional Equities

Jan-

09Ap

r-09

Jul-0

9Oc

t-09

Jan-

10Ap

r-10

Jul-1

0Oc

t-10

Jan-

11Ap

r-11

Jul-1

1Oc

t-11

Jan-

12Ap

r-12

Jul-1

2Oc

t-12

Jan-

13Ap

r-13

Jul-1

3Oc

t-13

Jan-

14Ap

r-14

Jul-1

4Oc

t-14

Jan-

15Ap

r-15

Jul-1

5Oc

t-15

Jan-

16

0

50,000

100,000

150,000

200,000

250,000Total passenger cars (units) 23%

Reduction in prediction error

(MAE) over traditional models

performs better near

Inflection points







Querimetrix—Business insightsONLINE EVALUATION CHANGING THE RULES

1

3

2

1. Evaluation

2. Car loans

3. Car exchange

Car finance—moves up the ladderSimultaneously assessing cars and loan options online

Car loan searches and car comparisons happening simultaneously

Google Correlate results for the keyword 'car loan' in India and US, 2004-15

Source: Google Correlate, Kotak Institutional Equities

• More loans to be sourced directly Loan origination costs may come down

• Dealer profitability may come down OEMs may need to increase incentive

• Car financing can be a differentiator Captive financing can innovate

3799

038

116

3824

238

368

3849

438

620

3874

638

872

3899

839

124

3925

039

376

3950

239

628

3975

439

880

4000

640

132

4025

840

384

4051

040

636

4076

240

888

4101

441

140

4126

641

392

4151

841

644

4177

041

896

4202

2

car loan (IN) compare cars (IN) emi calculator (IN)

Used car markets—influencing car salesThe effect on—1) New car sales and 2) Trade-ins

1 1+223%

yoy growth in used car queries

Online classifieds queries impede

new car sales

Trade-inan important

retention strategy

Online

SubstituteEntry-level car with a second-hand

car

Evaluation—landscape gets competitiveSearch-share for major car makers has dropped

2010 20150

20

40

60

80

100

120Maruti Suzuki Hyundai

Maruti Suzuki and Hyundai have both lost search-share since CY2010

Normalized query trends for Indian car OEMs, 2004-current


All options

are being evaluated

Evaluation—mindshare gets costlierAdvertising spends for the industry going up

Rising advertising spends indicate costlier mindshare

Estimated ad spends per car sale for OEMs in India, 2012-2015 (Rs/car)

Source: TAM, Kotak Institutional Equities

2012 2013 2014 201501,0002,0003,0004,0005,0006,0007,0008,0009,000

10,000

0

1,000

2,000

3,000

4,000

5,000

Maruti Udyog Ltd Hyundai Motor India LtdHonda Siel Cars India Ltd Total passenger carsAvg. TV unit costs (Rs/sec) (RHS) Avg. Print unit costs (Rs/sec) (RHS)

Notes:

(a) TAM data used for TV and Print advertising spends (discounted from rate card).

(b) Advertising spends are divided over total car and utility vehicle sales.

(c) TAM data available till August 31, 2015.

Quick take—Maruti SuzukiRetaining loyal customers and increasing reach

Howhas Maruti maintained

market-share?

Great strategy

is what helped retain market-share

Relating search queries and car salesThe relationship varies by OEM and models

CAR SALES

SEARCH QUERIES

• More model searches (i10 & i20), more sales

SIMPLE

SIMPLER

• More potential car buyers, more Maruti demand

QUERIMETRIX: WHATS NEXT?

Querimetrix: just warming up!Two-wheelers and Beauty are next on the cards

Dec-

04Ap

r-05

Aug-

05De

c-05

Apr-0

6Au

g-06

Dec-

06Ap

r-07

Jul-0

7No

v-07

Mar

-08

Jul-0

8No

v-08

Mar

-09

Jul-0

9No

v-09

Mar

-10

Jul-1

0Oc

t-10

Feb-

11Ju

n-11

Oct-1

1Fe

b-12

Jun-

12Oc

t-12

Feb-

13Ju

n-13

Oct-1

3Fe

b-14

Jun-

14Se

p-14

Jan-

15M

ay-1

5

0

10

20

30

40

50

60

Hero MotoCorp Bajaj Auto TVS Motor Companyhonda bikes Yamaha Motor Company Royal Enfield

Two wheelers do not mean Bajaj and Hero anymore!

Normalized 12-month rolling Google Trends for various two wheeler manufacturers

Source: TAM, Kotak Institutional Equities

THANK YOU

Option 1

presents

New Rules, New Tools!

Consumer Querimetrix

Presentations & Public Speaking