thinkvine boosting media value

© ThinkVine. All Rights Reserved.

A Breakthrough Approach to Marke0ng Mix Op0misa0on

in Today’s Consumer-‐driven World 4 October 2012


The Marketplace is Changing Rapidly

More fragmented

media consump0on

More diversity among

consumers

More marke0ng op0ons

More ways to buy


The Demand for Marke0ng Accountability is Rapidly Increasing

3

“Corporate marke,ng is the last bas,on of

unaccountable spending.” -‐ Google Chairman

Eric Schmidt


The Changes are Crea0ng New Challenges for Marketers

How do we best allocate the budget?

What size budget do we need?

How do we efficiently reach target segments?

What are the short-‐ and long-‐term effects? …


Marketers Are Struggling to Meet the Challenges

54% of marke9ng execs told BCG they do not have the tools required to make trade-‐offs

across media vehicles

65% of CMOs told IBM they are

“underprepared” for the growth in channel & device choices, 63% said it for demographic shiOs

9% of marke9ng execs informed the CMO Council that they have a highly evolved,

integrated marke9ng model

>50% picked “measurement, analysis and learning” and “integra9ng cross-‐channel marke9ng” as #1 and #2 challenges in a Unica poll


A Revolu0on in Marke0ng is Underway

6

From Annual to Agile Planning

From Mass to Engaged Branding & Communica0on

From Siloed to Integrated Execu0on

From Historical to Forward-‐Looking Analysis

From Cabals to Collabora0ves Organiza0onal Behaviour


A ThinkVine Marke0ng Mux Op0misa0on Success Story

•  Business Goals •  Acquire new customers with high life9me value •  Find the op9mal spend level and the ROI of tac9cs

•  Discoveries •  Using ThinkVine soOware the company discovered

•  A significant shiO from tradi9onal media to digital would meet goals •  The effect of each marke9ng tac9c on acquisi9on volume and its ROI •  The likely effects of new marke9ng tac9cs at scale •  The shiO to digital would affect brand awareness in future years

•  Ac0ons •  Made shiO to digital, beat customer acquisi9on targets in first year •  Adjusted mix in second year to beXer balance short-‐ and long-‐term •  Began integra9ng detailed digital data into planning

7

Financial Services


Op0misa0on Requirements •  Agile •  Accurate, objec9ve historical analysis and, more importantly, forecasts of “Boardroom Metrics” •  Volume, Sales, ROI, CLTV, etc. •  AXribu9on

•  Mul9dimensional views •  All marke9ng tac9cs •  Consumer segments •  Geographies

•  Reliable, ac9onable informa9on •  Collabora9ve process •  Flexible and up-‐to-‐date models •  Low total cost of ownership

Marke0ng Mix Today

Analyse impact across channels

8

Forecast marke9ng outcomes

What people want

Measure Marke9ng ROI

Value people receive

Achieving goals with rela9vely low investment

Sharing business insights across the enterprise

Analyse the past, present and future


Marke0ng Mix Op0misa0on Op0ons

Regression-‐based Mix Agent-‐based Mix

9



Regression-‐based Mix Mathema9cal approach. Consultant/agency or internal teams do as a project. Technical capabili9es and marke9ng exper9se vary by company and analyst

Historical analysis with some granularity “teased out” of the top-‐down model. Only “sees” large amounts and variances

Forecasts extend history. Validity is bounded by % change. Granularity is oOen constrained. No consumers or new tac9cs in the forecast

Forecasts are typically one number. Accuracy has a short half-‐life

Fixed form equa9ons, changes are effec9vely new projects

No collabora9on capability. No saved plans, typically no saved forecasts

5-‐6 figure project price plus add-‐on for simulators and updates

10



Regression-‐based Mix Agent-‐based Mix (ThinkVine) Mathema9cal approach. Consultant/agency or internal teams do as a project. Technical capabili9es and marke9ng exper9se vary by company and analyst

Complex systems approach newly applied to marke9ng by using data about consumer behaviour. The soOware’s simula9on capability and the marke9ng exper9se vary by company

Historical analysis with some granularity “teased out” of the top-‐down model. Only “sees” large amounts and variances

Historical analysis with significant granularity for all types of marke9ng developed boXom-‐up, including on-‐line and off-‐line aXribu9on

Forecasts extend history. Validity is bounded by % change. Granularity is oOen constrained. No consumers or new tac9cs in the forecast

Forecasts via forward-‐looking scenarios that make extensive use of custom consumer response data for all marke9ng types and levels

Forecasts are typically one number. Accuracy has a short half-‐life

Forecasts are mul9-‐dimensional with significant granularity. High accuracy over long periods

Fixed form equa9ons, changes are effec9vely new projects

Ongoing data updates, periodic recalibra9on if needed as part of subscrip9on

No collabora9on capability. No saved plans, typically no saved forecasts

Built-‐in collabora9on func9onality. Plans and forecasts are saved in the soOware

5-‐6 figure project price plus add-‐on for simulators and updates

5-‐6 figure annual subscrip9on fee

11


Organiza0onal Profiles Influence the Decision

Regression-‐based Mix •  Familiarity with regression is important to organiza9on

•  Marketplace is changing slowly •  Only use a few types of marke9ng, don’t plan to try new things

•  Limited or no segmenta9on •  Minimal compe99on •  Plans are made once per year and rarely adjusted

•  Collabora9on with &/or second opinion of agency isn’t important

•  Okay with 9me lag for new projects to get updates

•  Internal or low-‐cost source

Agent-‐based Mix •  Organiza9on is open to change, uses analy9cs for compe99ve advantage

•  Marketplace is changing rapidly •  Use many types of marke9ng, experimen9ng with new things

•  Segmenta9on is important •  Ac9ve compe99on •  Plans are updated throughout the year based on results

•  Collabora9on with &/or second opinion of agency is important

•  Want to run scenarios themselves •  Want to synchronize and integrate planning and execu9on

12


Mix and Digital Cross-‐Channel Adribu0on Have Different Roles

13

ThinkVine MMO Digital X-‐Channel Adribu0on

Focus Strategic planning Tac9cal execu9on

Objec0ve Marke9ng budget op9miza9on Campaign conversion op9miza9on

Emphasis Achieve key goals Realize incremental gains

Method Effec9veness via behaviour simula9on Efficiency via itera9ve execu9on

Forecasts Short-‐ and long-‐term sales forecasts 2-‐8 week campaign forecast

Advantages Breadth across all tac9cs, consumer segments and channels Depth within digital

ThinkVine can integrate detailed digital (and social, etc.) data to create a “virtuous circle” that improves both planning and execu,on


A ThinkVine Marke0ng Mix Op0misa0on Success Story

•  Business Goals •  Increase market share in very compe99ve category •  Reverse decline in brand awareness

•  Discoveries •  Using ThinkVine soOware the company discovered

•  Erosion in brand recogni9on was crea9ng an ever-‐higher hill to climb •  Compe9tor’s television blitz had further eroded brand recogni9on •  The right budget level, mix and 9ming could quickly pay back

•  Ac0ons •  Won Board approval to increase and shiO budget •  Increased television adver9sing focused on innova9on with corresponding digital content to facilitate path-‐to-‐ac9on

•  Achieved objec9ves within six months, got “second tranche” funding •  Began digging deeper into the effects of pricing in the mix

14

Consumer Services


ThinkVine Marke0ng Mix Op0misa0on Sogware

Agent-‐based Modelling (ABM) Overview

15

Is agent-‐based modelling new?

ABM has been around since the 1940’s with roots in the hard sciences

Very good at understanding how many moving parts influence an outcome

Used extensively by Boeing, WHO, DoT

“Regression wasn’t invented for marke,ng mix; it was applied to it.”


ThinkVine Marke0ng Mix Op0misa0on Sogware

Agent-‐based Modelling (ABM) Overview

16

Why hasn’t agent-‐based modelling been applied to marke9ng un9l now?

Lack of need – yesterday’s marketplace was less complex, more homogeneous

Lack of compu9ng power – simula9ons are computa9onally intensive to run

Lack of data & applied IP – about consumer response to marke9ng

“ThinkVine’s marke,ng mix op,misa,on soJware harnesses the advantages and advancements of ABM to address the growing complexity in today’s marke,ng landscape.”


ThinkVine’s Sogware: “Sim City for CMOs”

17

Marke0ng Ac0vi0es Sales

•  ThinkVine’s marke9ng mix op9misa9on soOware enables you to simulate the results of your marke9ng plans in a marketplace where a representa9ve popula9on of simulated people consume media and purchase products the way real people do

•  It combines demographic, consumer behavior and market data with your data and advanced analy9cs to deliver the forward-‐looking informa9on you need to maximize results – accurate forecasts with rich, mul9dimensional informa9on by consumer group, type of marke9ng, product, channel and geography


Find the Op0mal Investment Level to Meet Your Objec0ves

18

•  The soOware enables your team to easily run “What If?” scenarios. You can find the op9mal investment levels to meet various strategic and tac9cal objec9ves.

ROI and Gross Margin by Total Marke0ng Investment

2011 $100 MM Spent

What If? $145 MM Budget

What If? $300 MM Budget

Max ROI

Max Margin


•  The soOware enables fast, objec9ve comparisons of marke9ng plan alterna9ves. You can vary spending levels, 9ming, mix of tac9cs and consumer groups targeted.

Iden0fy the Marke0ng Plan Most Likely to Deliver the Best Results

19

2011 $180 MM Spent

170 173

168

175 178

185

Units (M

M)

2012

$180 MM Budget

Same Mix as 2011

2012 $180 MM Budget

Shig $5 MM Print to TV

2012 $175 MM Budget

Cut $5 MM in Print

2012 $185 MM Budget

Shig $5 MM Print to Digital Add $5 MM TV

+4.7%

+1.8% +2.9%

+8.8%

-‐1.2%

Comparison of 2012 Marke0ng Plan Alterna0ves

2012 $180 MM Budget

Shig $5 MM Print to Digital


Discover Which Marke0ng Tac0cs Will Yield High ROI

20

•  The soOware analyses and predicts how marke9ng tac9cs influence purchases. •  You see granular informa9on about how each performs in the context of the overall plan.

ROI by Type of Marke0ng Ac0vity


Learn What Works Best by Consumer Segment

0% 10% 20% 30% 40% 50% 60%

Other <30

<30 Affluent

50+ Other

50+ Affluent

TV

Radio

Print

Digital Ads

Search

Mobile

21

•  The soOware reveals how well marke9ng tac9cs work alone and in combina9on by consumer group. You can tune your mix to efficiently reach segments you target.

Marke0ng Tac0c Effec0veness By Consumer Group


Understand the Short-‐ and Long-‐Term Value of Marke0ng

22

74% 50% 38%

26% 28% 18%

22% 25%

19%

0

20

40

60

80

100

120

140

160

2012 2013 2014

2014 Plan

2013 Plan

2012 Plan

Pre-‐2012 Base

•  The soOware forecasts current and future year effects of marke9ng plans. You can see the cumula9ve value of new marke9ng strategies and ini9a9ves.

Units (M

M)

Effect of Proposed Marke0ng Plans By Year

81% Base

78% Base

“Brand value”


Forecast Results of New Marke0ng Ini0a0ves

23

2012 Original 2012 Add Mobile % Change

Total Units (‘000s) 36,928 38,109 3%

Digital 22%

Mass Media 67%

Promo0ons 11%

Digital 22%

Mass Media 61%

Promo0ons 11%

Mobile 6%

Planned Marke0ng Mix Proposed Mix with Mobile Added

Unit Sales Forecast

•  The soOware provides informa9on about the likely value of all marke9ng tac9cs, including tac9cs not used before. You understand the poten9al before you invest.


ThinkVine Marke0ng Mix Op0misa0on Sogware Our Approach

24

ThinkVine replicates reality by crea9ng a custom, simulated marketplace in which you can measure purchase behaviour across consumers and by marke9ng ac9vity.

Build Agent Popula0on Consumer agents are configured to be highly representa9ve of a popula9on using demographic and media consump9on data.

Add Customer Data The agents’ behaviour is informed by customer-‐supplied data about the category, the product, pricing, marke9ng ac9vity, etc.

Create Market Model ThinkVine’s experts develop a custom model in which the agents respond to marke9ng ac9vity and make purchases within the category in a realis9c way.

Ongoing Planning The soOware enables marketers to quickly and easily run scenarios in a frequently updated marketplace and get a rich, objec9ve set of decision-‐making informa9on.

1

2

3

4


Building an Agent Popula0on Demographics

25

Sample Consumer Male 47 Years Old Caucasian Income, $110,000 / yr.

•  We mathema9cally recreate a representa9ve popula9on of target markets by sta9s9cally drawing from empirical distribu9ons of many variables

•  Targe9ng variables include geography, income, age, ethnicity, etc.


# of minutes shopping online per day

# of minutes watching television per day

# of minutes using social media per day

Building an Agent Popula0on Media Consump0on

26

•  We assign the agents media consump9on behavior by tapping a variety of widely-‐accepted sources of consumer behaviour data

•  We use patent-‐pending IP to sample from these known distribu9ons and assign media exposure probabili9es

•  As a result, some agents watch a lot of television (and have higher exposure probabili9es) while others will watch liXle to no television (and have lower probabili9es)


Building an Agent Popula0on Consumer Agents

27

•  The soOware now contains a marketplace of agents that is representa9ve of the consumers in your marketplace

•  It also contains marke9ng science-‐based rules about response and purchase behaviour


Adding Customer Data ThinkVine Marke0ng Mix Op0miza0on Sogware: Your Data

Marke0ng plans Media, promo9on, marke9ng, digital Sales data Syndicated data, internal sales data Market research data Segmenta9on, price elas9city, exogenous factors, addi9onal general and demographic-‐based media consump9on data, etc.

28

Required Data

Enhance & Improve Your Results


Adding Customer Data Purchasing Behavior

29

•  We use the customer data to configure the agents so their purchase behaviour matches how consumers actually purchase within the specific category

Brand Awareness

Category Purchase Frequency

Purchase Probability

~90% 4

~20%


Crea0ng the Market Model

A “Bodom-‐up” Approach That Predicts More Accurately

30

Training the system ThinkVine’s analysts develop and calibrate a custom model that recreates past sales from the “boXom up” by establishing consumer-‐centric rules that fit historical results.

Proving the system Once the system is trained, we validate it by comparing the most recent 6 months of actual sales (“hold out” data not used to train the system) to a simula9on of that 6 month period.

Using the system Once the model has been calibrated & validated, it is ready to accurately forecast future sales and provide rich decision support informa9on for each marke9ng plan and “what if” scenario.

Train Prove Use

Calibra9on using demographic, consump9on, plan, research & sales data Validate with hold out data Forecasts

2 Years Sales History 1-‐2 Quarters Future

Simulated Actual

1

2

3


ThinkVine and Partners Deliver Value Quickly and Over Time

31

Configura0on & Ini0al Guidance

•  Analysis of recent marke9ng ac9vity

•  Recommenda0ons for immediate changes in the marke9ng mix

•  Accurate forecasts

Agile, objec0ve marke0ng planning & op0misa0on

•  Respond quickly to market and budget changes by running “What If?” analyses of marke9ng plans

•  See forecasts with rich informa0on by consumer group, marke9ng tac9c, channels and geography

•  Iden0fy new opportuni0es for growth and savings

•  See the effects of media you haven’t tried before

•  Create effec0ve, efficient integrated plans across all marke9ng ac9vi9es

•  Receive expert ongoing marke0ng support


Sample Screens: Compare Plans, Review by Tac0c, Segment, Etc.

33

High-‐level By group & year

Long-‐term effects of plans

Compare ROI of plans, filter results by 9me, segment, etc.

“Due to” and ROI by tac9c and segment


A ThinkVine Marke0ng Mix Op0misa0on Success Story

•  Business Goals •  Increase market share while maintaining margins •  Improve overall marke9ng ROI

•  Discoveries •  Based on recent success with a specific consumer group, the agency recommended doubling the targeted radio ads at mid-‐year

•  Using ThinkVine soOware the company discovered •  Ads drive traffic to store, but in-‐store price drives purchase decisions •  The marginal return from more ads to the group would be very low •  There was a large consumer group with more upside poten9al

•  Ac0ons •  Adjusted its marke9ng mix to achieve its business goals •  Now quan9fies all proposed changes via planning soOware

34

Consumer Products


ThinkVine's marke9ng mix op9misa9on soOware provides B2C marketers with the forward-‐looking informa9on they need to

get the maximum return across all of their marke9ng investments.

See a 2 minute video overview at

www.thinkvine.com or call

+1 513 842 5900

thinkvine boosting media value

Technology

new marke9ng tac9cs

marke9ng exper9se

types of marke9ng

marke9ng types

channel marke9ng

marke9ng execs

technical marke9ng

saved forecasts forecasts