Four As Model Multivariate Solutions Multivariate Solutions June - - PowerPoint PPT Presentation

Four A’s Model

Multivariate Solutions Multivariate Solutions June 2005 June 2005

• To examine four key measures in the

study to determine key segments

– Satisfaction with The Company – Would Recommend The Company – Likelihood to Remain with The Company – Magazine Purchase Intent

• To Determine key discriminators for each

segment within the Orange (The Company) and Blue (Triple Artists) groups.

Background and Objectives

• Adorer

– High scores on all measures and will recommend The Company

• Adopter

– High scores or will recommend The Company on most, but not all, key measures

• Acceptor

– High scores on half or fewer key measures

• Rejectors

– No high scores on any key measures

The Key Segments Model The Key Segments Model -

• Snapshot

Snapshot

The Key Segments Model The Key Segments Model -

• Frequencies

Frequencies

Percentage Total Recipients Non- Repicents O range Database Blue Database

A dorer 21 23 19 22 20 A dapters 36 34 37 38 33 A cceptors 31 31 30 30 32 Rejectors 13 12 14 10 15

The Key Segments Model The Key Segments Model -

• Frequencies

Frequencies

0.9 1.1 1.1 1.9 Total Purchases 1% 2% 4% 15% Purchase a The Company Gift Card 4% 9% 10% 18% Subscribe to Roadside Assistance 15% 22% 23% 39% Download a Ringtone 8% 9% 9% 12% Set Up International Calling Services 15% 14% 15% 25% Browse the Wireless Internet 29% 29% 29% 50% Use Text Messaging 23% 19% 21% 27% Purchase a Cell Phone 1.8 3.7 4.3 9.1 Purchase/Read Interest Magazine 0% 54% 93% 100% Will Not Leave The Company 0% 68% 94% 100% Would Recommend to a Friend 4.1 6.5 8.5 9.0 The Company Satisfaction Rejector Acceptor Adopter Adorer

Key Measures

The Key Segments Model The Key Segments Model -

• Frequencies

Frequencies

0.9 1.1 1.1 1.0 1.2 1.1 2.1 1.7 Total Purchases 1% 0% 3% 1% 6% 3% 16% 14% Purchase a The Company Gift Card 3% 7% 8% 11% 9% 10% 17% 19% Subscribe to Roadside Assistance 12% 20% 23% 21% 23% 24% 44% 35% Download a Ringtone 8% 8% 10% 7% 13% 6% 18% 6% Set Up International Calling Services 13% 16% 15% 14% 17% 13% 26% 25% Browse the Wireless Internet 28% 32% 35% 23% 31% 28% 58% 43% Use Text Messaging 21% 26% 18% 21% 21% 21% 29% 25% Purchase a Cell Phone 1.6 2.2 3.9 3.4 4.4 4.1 9.2 9.1 Purchase/Read Interest Magazine 0% 0% 54% 54% 94% 92% 100% 100% Will Not Leave The Company 0% 0% 65% 72% 90% 97% 100% 100% Would Recommend to a Friend 3.8 4.5 6.4 6.5 8.4 8.6 9.1 9.0 The Company Satisfaction Blue Orange Blue Orange Blue Orange Blue Orange Rejector Acceptor Adopter Adorer

Key Measures

Discriminant Analysis

• Discriminant analysis generates a function based on linear combinations of the

predictor variables (e.g. Off Peak Minutes or Shared Plan) that provide the best explanation between known groups. Specifically, the functions are generated from the key measures clusters.

• The following graphs display the relative impact of each of the included variables had
• n the segments. Values shown are coefficients that explain the descriptive impact

each variable had on the segment. They are relative.

– That is, if ‘Total Month’ has a coefficient of, say, .96, and ARPU had a coefficient of .46, ‘Total Month’ has twice the impact on the equation. – Variables displayed showed a discernible impact.

• These are interpreted as key discriminators for the segments.
• Predictive Value

– It is used in situations where you want to build a predictive model of segment membership based on observed data — off-peak minutes, share plan, etc. – The analysis produces a linear equation of included variables that can be used to forecast group membership for future respondents. – For example, The Company would like to find ‘Adorers’ from their main database. Using the discrimininant function they can calculate a scores for each mobile number, then sort the scores from high to low. Those scores at the top are far more likely to be the target group, ‘Adorers’.

0.96 0.92 0.46 0.26 0.25 0.19 0.08 0.02

0.0 0.2 0.4 0.6 0.8 1.0

T o t al M o nt h Of f p eak M inut es A R PU A ccess R evenue R o ad Sid e Service SM S so licit Share Plan Peak M inut es

Key Segments Key Segments – – Discriminant Discriminant Analysis Analysis

ORANGE DATABASE ORANGE DATABASE

The Key Segments Model The Key Segments Model – – Discriminant Discriminant Variables Variables ORANGE DATABASE ORANGE DATABASE

35.8 35.4 32.9 33.2 Access Revenue 421.1 375.6 375.7 375.9 Peak Minutes 200.7 202.2 209.4 239.0 Offpeak Minutes 49.9 50.2 48.2 46.7 ARPU 708.1 699.8 724.1 740.1 Total Month 58% 61% 65% 59% SMS Solicitation 26% 18% 22% 22% Roadside Assistance 62% 49% 48% 49% Share Plan Rejector Acceptor Adopter Adorer

Discriminant Measures

1.03 0.50 0.44 0.40 0.35 0.24 0.02

0.0 0.2 0.4 0.6 0.8 1.0 1.2

A ccess R evenue A verag e R evenue Int ernat io nal Plan A verag e F eat ure A mo unt Share Plan R ing t o ne T o t al A verag e M o nt hly U sag e

Key Segments Key Segments – – Discriminant Discriminant Analysis Analysis

BLUE DATABASE BLUE DATABASE

The Key Segments Model The Key Segments Model – – Discriminant Discriminant Variables Variables BLUE DATABASE BLUE DATABASE

809.6 775.1 772.3 810.5 Average Monthly Usage 42.4 42.3 41.5 39.3 Access Revenue 4.7 3.3 2.9 2.5 Ringtone Total 3.4 3.1 4.1 3.7 Average Feature Amount 61.0 57.0 56.1 52.9 Average Revenue 9% 11% 14% 9% International Plan 42% 45% 44% 47% Share Plan Rejector Acceptor Adopter Adorer

Discriminant Measures