Lezione in Inglese al Politecnico di Milano al Master internazionale in Marketing del 2015. Quest'anno (2016), la lezione sul CRM sarà il prossimo Febbraio 24
2. 2
The marketing process (1/2)
Strategic objectives
Targeting
Choice of market segment/segments
Positioning
Definition of the distinctive aspects on whihc we would like to be
perceived as different and preferable
Value proposition
Definition of the benefit/value we intend to transmit to the customer
Marketing Mix
Planning the marketing lever/action plan
Realization and control
Understanding of the external environment
-Breadth, structure, market trends
-Segmentation
-Competition
-Customer needs and expectations
-Influence of external factors
Understanding of our own position in the market
-Comparison between our market services and those of the
competition
-Diagnosis of the reasons for the diversity of services
Our strengths and weaknesses v our competitors
OPERATIONALSTRATEGIC
Opportunities and threats
3. 3
The marketing process (2/2)
Strategic objectives
Targeting
Choice of market segment/segments
Positioning
Definition of the distinctive aspects on whihc we would like to be
perceived as different and preferable
Value proposition
Definition of the benefit/value we intend to transmit to the customer
Marketing Mix
Planning the marketing lever/action plan
Realization and control
Understanding of the external environment
-Breadth, structure, market trends
-Segmentation
-Competition
-Customer needs and expectations
-Influence of external factors
Opportunities and threats
Understanding of our own position in the market
-Comparison between our market services and those of the
competition
-Diagnosis of the reasons for the diversity of services
Our strengths and weaknesses v our competitors
OPERATIONALSTRATEGIC
4. 4
Market analysis logic – a new marketing approach
The marketing approach considers three dimensions for
market analysis:
1. Macro-environment analysis
2. Demand analysis
A. From a “macro” viewpoint:
I. Definition and qualification of the market (boundaries, breadth, etc.)
II. Measurement of the market (estimate and forecast of demand)
B. From a “micro” viewpoint:
I. Analysis of market needs
II. Analysis of buying behaviour
3. Offering analysis
A. Analysis of the competitive system
5. 5
Demand analysis: a reference framework
Objectives:
To analyse the target market to evaluate its attractiveness,
the decision to enter and the marketing strategy.
1. Evaluation of the marketability of new products.
2. Decision to exit a market.
Tasks:
1. Analyse the needs and wants of existing and potential
customers
2. Qualify and estimate the reference market Macro demand
analysis
3. Analyse customers’ buying behaviour Micro demand analysis
9. 9
Political analysis
Political / legal factors
Licenses (e.g. bakeries)
Free trade agreements (e.g. WTO, NAFTA, EU ecc.)
Legal environment (e.g. consumption taxes vs. production taxes)
Country V.A.T. Corporate tax rate
Argentina 21,0 %
Bulgaria 20,0 %
Cyprus 17,0 % 10%
Denmark 25,0 %
France 19,6 % 33,22%
Germany 19,0 % 29,48%
Italy 22,0 % 31,4%
UK 20%
Romania 24,0%
Russia 18,0 % 20%
Switzerland 8% 21,7%
10. 10
Socio-Economic, Technological & Environmental
analysis
Economic factors Social factors
• Conjunctural impact on the demand
• Average purchase power
• Distribution of wealth
• Interest rate
• Health consciousness
• Population growth rate
• Age distribution
• Career attitudes
• Emphasis on safety
Technological factors Enviornmental factors
• Scientific/technological discoveries
• Technological infrastructures
• Weather
• Climate and climate change (e.g. tourism)
• Environmental awareness
13. 13
Example: OECD Countries
Organisation for Economic Co-
operation and Development.
OECD
Country
AUSTRALIA
AUSTRIA
BELGIUM
CANADA
CHILE
CZECH REPUBLIC
DENMARK
ESTONIA
FINLAND
FRANCE
GERMANY
GREECE
HUNGARY
ICELAND
IRELAND
ISRAEL
ITALY
JAPAN
KOREA
LUXEMBOURG
MEXICO
NETHERLANDS
NEW ZEALAND
NORWAY
POLAND
PORTUGAL
SLOVAK REPUBLIC
SLOVENIA
SPAIN
SWEDEN
SWITZERLAND
TURKEY
UNITED KINGDOM
UNITED STATES
13
On 14 December 1960, 20 countries originally
signed the Convention on the Organisation for
Economic Co-operation and Development. Since
then, 14 countries have become members of the
Organisation.
Here is a list of the current Member countries of the
Organisation and the dates on which they deposited
their instruments of ratification.
14. 14
Innovative countries
14
The Global Innovation
Index (GII) is a
recognition of the key
role that innovation
serves as a driver of
economic growth and
prosperity.
20. 20
Market qualification
Potential market: market that is achieved when the
marketing effort is infinite
Qualified market: market with the requisites to be
available for the purchase (e.g. 18+ y.o. – or 21 y.o. in some
countries – for some products; driving license, etc.)
Available market: qualified + able to spend enough to
buy the product or interested enough in the category
Served market: available market that is reached by the
marketing effort of the companies
Penetrated market: market that has already purchased
products/services in the category
22. 22
Potential Market of Cloudea
Dimensione azienda
Somma di
IMPRESE/ISTITUZIONI
Somma di IMP.
INFORMATIZZATE
Somma
di IB FY
2009
Somma di
ADDETTI
Core MM (50-249 PCs) 10.904 10.900 972.884 2.117.021
Core SB (5-24 PCs) 305.733 192.066 1.817.466 3.573.000
Home Business/Low SB (1-4
PCs)
4.002.652 2.147.879 4.260.164 7.782.820
Low MM (25-49 PCs) 13.439 13.327 450.897 925.012
Totale complessivo 4.332.728 2.364.172 7.501.411 14.397.853
Settore
Penetrazione Iaas
Pubblico
Penetrazione SaaS
Pubblico
Servizi (Editoria, Media, ICT,TLC) 15% 63%
Industrial 49% 64%
PA 12% 59%
Utility/Oil & Gas 27% 47%
Bancario, Assicurativo 21% 36%
GDO/Vendita al dettaglio 30% 40%
23. 23
Demand estimate and forecast
Processes and methods for obtaining a
qualitative and quantitative valuation of
demand
The choice of the right model depends on
the specific objectives and on the
availability of quantitative data to base
the esteem
24. 24
Choice of the market forecasting method
1. Based on availabile data
A. Quantitative data available? Quantitative methods
B. Quantitative data not available? Qualitative
methods
2. Based on the objective. Three main objectives:
A. Estimating market size of an existing product?
B. Estimating the market demand after a new product
launch?
C. Estimating the market demand precisely in a time-
lapse?
25. 25
Forecasting methods
Objective
Estimate
market size
Estimate demand for
new products
Estimate demand for
established products
Quantitative
• Coefficient
method
• Diffusion models
• Analytical /
epidemiological
• Time series
• Linear Regression
Qualitative
•“Desk”
methods
• Gaussian
•Adoption models
•Market tests
•Market research
26. 26
Forecasting methods
Objective
Estimate
market size
Estimate demand for
new products
Estimate demand for
established products
Quantitative
• Coefficient
method
• Diffusion models
• Analytical /
epidemiological
• Time series
• Linear Regression
Qualitative
•“Desk”
methods
• Gaussian
•Adoption models
•Market tests
•Market research
27. 27
Coefficient method
What are coefficients?
Parameters representing a past experience or
a future expectation on the evolution of a
phenomenon
- Consuming good
- Industrial good
- Investment good
Coefficient models vary according to:
- Product type
- Repurchase rate
27
Estimate market size
28. 28
Fast moving consumer goods and services, not tied to the
use of equipment (e.g. preserves, personal or home care
products, etc.)
N x % x Cu
n
N
Potential market size
% effective users of
the category
(COVERAGE)
Individual rate of
consumption
(PENETRATION)
Market research and observation
Q=
Market size (n°
units)
28
Estimate market size
29. 29Marketing Multicanale - a.a. 2009/10 - prof.Giuliano Noci 26
Quarterly demand of throw-away razor blades in the
Italian market
25.107.509 x x 415 %
Number of Italian males
over 14 y.o.
% of throw-away razor
blades users
Number of blades
used quarterly on
average
Q=
Market size (n°
units)
29
= 15.064.505
Estimate market size: an example
30. 30
Consuming goods tied to the use of specific equipment
N x % x Ta x ct
na
N
% owning the
equipment
Use rate of the
equipment
Number of units of goods
used per single use
(TECHNICAL COEFFICIENT)
Market research and
observation
Product characteristic
Q=
Potential
market size
(n° units)
30
Estimate market size
31. 31Marketing Multicanale - a.a. 2009/10 - prof.Giuliano Noci 28
Monthly demand for dishwashing tabs
20mln x % x 25 x 135,4
Penetration of
dishwashing
machines
Monthly # of
washing
Number of tabs per washing
cycle
Q =
Number of Italian
households
31
= 177.000.000
Estimate market size: an example
32. 32
Desk Research
Desk Research is the research technique which is mainly acquired by
sitting at a desk. Desk research is basically involved in collecting data
from existing resources hence it is often considered a low cost
technique as compared to field research, as the main cost is involved
in executive’s time, telephone charges and directories.
Internal Desk Research
External Desk Research
Online Desk Research
Government published data
Customer desk research
Analyst published data (free or paid)
33. 33
Forecasting methods
Objective
Estimate
market size
Estimate demand for
new products
Estimate demand for
established products
Quantitative
• Coefficient
method
• Diffusion models
• Analytical /
epidemiological
• Time series
• Linear Regression
Qualitative
•“Desk”
methods
• Gaussian
•Adoption models
•Market tests
•Market research
34. 34
Epidemiologic models
Two kinds of users
Purchasers (infected)
Non-purchasers (not enfected)
Purchasers “infect” non-purchasers
Like in an epidemy…
The higher the number of purchasers in the time t, say Q(t), the
higher the probability of infection
34
Demand forecasting
Wikipedia : Journal o Clinical Epidemiology
35. 3535
Infection coefficient (or infection probability) y is
distributed like a logistic curve
t
1/2
1
y
t = -K/r
r = diffusion rate of innovation
K = starting condition
Demand forecasting
36. 3636
Coefficient y depends on
Q(t)
The number of purchasers in t
Q*
The potential market
y=
Q(t)
Q*
____
Demand forecasting
37. 3737
Multiplying Q* times and deriving, we get the
purchasers in the exact time t
q(t)
t = -K/r
q(t)= y’(t)Q* x
t
Demand forecasting
38. 3838
Customers have different attitudes towards first purchase of a
new product
• Precursors (buy right because the product is new)
• Innovators (recognize value-for-money and usefulness)
• innovative majority (deliberation, prudence)
• conservative majority (skepticism)
• conservatives, or last buyers (tradition)
• unyielding
Demand forecasting
40. 40
Forrester Research on Purchasing influence
In the definition of
the process we
always remember
what influences
people to buy in
different countries
because it might be
too different from
one area to another
43. 43
Rogers’ Curve Examples
Today’s social and
cloud services (2011)
How political culture
influence the high tech
buying attitude and
how innovators are
gaining
46. 4646
Pros Cons
Experts •External (multiple) viewpoint
•Voice of experience
•Statistically insignificant
•Potentially costly
Sales force •Cheap and timely
•Direct contact with customers
•Statistically insignificant
•Only internal
Customers •Statistically significant
(if quantitative)
•Unfiltered
•Complex
•Costly
•Long
Demand forecasting
47. 47
Forecasting methods
Objective
Estimate
market size
Estimate demand for
new products
Estimate demand for
established products
Quantitative
• Coefficient
method
• Diffusion models
• Analytical /
epidemiological
• Time series
• Linear Regression
Qualitative
•“Desk”
methods
• Gaussian
•Adoption models
•Market tests
•Market research
48. 4848
A. Past demand
B. Factors influencing
demand
Time series
Causal models
Established products
Quantitative approaches Forecast based on:
Demand forecasting
51. 51
Demand forecasting
A time series is made up of a combination of 4 “movements”:
Trends (T): the underlying tendency that affects the long term.
Cyclicity (C): the tendency of the business cycle that characterises the
course of the economy in a particular sector in the medium-long term (3 –
7 years).
Seasonality (S): involves shorter periods and refers to patterns that are
repeated in more or less the same course during the subsequent
corresponding periods (years, months, days).
Randomness (E): refers to the variations that cannot be explained by the
previous characteristics random error.
52. 52
Multiplicative model
Analyzing a time series it is possible to estimate the effect of the four
components
• Trend
• Moving average (MA)
• Weighted MA
• Exponential smoothing
• Cyclicity and Seasonality
• Ratio approach
• Randomness
• Inestimable by definition analyzing the time series, may be estimated
through contextual analysis (e.g. exceptional marketing effort by companies,
etc.)
ESCTD ***
Demand forecasting
53. 53
Moving average
Allows an approximate evaluation of the trend: reducing the seasonal
and random variations
where:
Aj= demand of period j
N = number of periods on which the average is calculated
The moving average eliminates the passing fluctuations giving rise to a
rounding-off effect.
The number of periods on which the average is calculated influences
the result (the higher N is, the more the rounding-off effect).
If N is the seasonality period, the moving average has a
de-seasonalising effect.
Trend = M(t)/M(t-1)
Trend analysis
54. 54
Trend analysis (2/3)
Weighted moving average
Enables one to give more weight to more recent data and
decreasing weight to others according to preference
where:
Pj= weight of value in the period j
t
1Nt
j
1Nt1Nt1t 1tt t
t
p
ApApAp
MP
.......
55. 55
Cyclicity and seasonality
Multiplicative model
Seasonality index (monthly, weekly or daily):
More precise estimate: calculating the seasonality index in
different intervals and averaging
ESCTD ***
T
ED
SC
*
*
A
D
S t
t
Where:
- D(t) = demand in the interval t
- A = average demand in a not-seasonalised interval
56. 56
Cyclicity and seasonality
For cyclicity the approach is substantially the same, but:
t is generally a year
A is the average yearly demand (MA) divided by a trend
coefficient
A
D
C t
t
57. 57
Using time series for demand forecasting
Demand forecasting through time series
hypothesizes that demand in the interval (t+1) is
related somehow to the demand in the interval (t)
The function f is the combined effect of the time
series components
)( tt DfD 1
58. 58
Time series decomposition method – example (1/2)
Consider a company that produces hollow glass; the table below
shows sales for 2008 and some elements that characterise the past
series
Forecast sales for the month of June 2009
2008 sales in tonnes 492,787
Annual growth trend (average last 2 years) +4.2%
Cyclical component Irrelevant
Seasonal component (June/July/August) +15%
Random component (environmental sensitivity + others) +2%
59. 59
Time series decomposition method – example (2/2)
Consider 2008: let’s say demand is constant in all months
except June, July and August, when there is an increase of 15%
(seasonality of the product)
9X + 3*1,15X = 492.787 ton
X= 39.581 monthly sales in 2008 (except in June, July and August)
Forecast for 2009, monthly sales:
Every month except, June, July and August:
T + E = 39.581*1,042*1,02 = 42.068 ton
June, July and August:
other months’ demand + S = 42.068*1,15 = 48.970 ton
Implicit hypotheses: T, S and E are constant in time
60. 60
Forecasting with moving average
Often the multiplicative approach is complex
• Need to estimate the cycles
• Need of data over a long time set
A cheaper approach: moving average to forecast
demand
Simplest case: no underlying trend
Ft+1 (next period demand) is generally equal to
the average
When a increasing/decreasing trend exists use
of exponential smoothing to emphasize the trend
M
AAA
F t
1Nt1tt
1t
N
.......
Stock exchange
indexes analisys
simple
40 periods
61. 61
Example (1/2)
Based on historic data, make a forecast for
December using the moving average
MONTH PERIOD EFFECTIVE
DEMAND
3 MONTH
FORECAST WITH
MOVING
AVERAGE
2 MONTH
FORECAST WITH
MOVING
AVERAGE
JANUARY 1 230
FEBRUARY 2 135
MARCH 3 210 183
APRIL 4 238 192 173
MAY 5 48 194 224
JUNE 6 225 165 143
JULY 7 155 170 137
AUGUST 8 30 143 190
SEPTEMBER 9 320 137 93
OCTOBER 10 250 168 175
NOVEMBER 11 210 200 285
DECEMBER 12 260 230
63. 6363
Causal methods
Data on the factors
influencing demand
Causal models
(e.g. regressive
models)
Demand forecasting
Forecasting
64. 6464
Examples of regressive causal models
• Linear regression models
• Logit models (logistic regression)
etc.
Demand forecasting
Stock exchange
values analisys
http://www-
bcf.usc.edu/~gareth/ISL/ISLR
%20Fourth%20Printing.pdf
65. 65
Linear regression models
Q = k + a*F1 + b*F2 + …. + ε
Q = demand Fi = explanatory factors
K = constant (known term) ε = random error
(unexplainable)
Aim: to estimate the coefficients (a,b,c,…).
Generally OLS (Ordinary Least Square) are used to
obtain:
1. Estimates of the coefficients and of the constant
(correct a statistical substance)
2. Reliability indicators of the method:
R2: % of the variability of the dependent variable
explained by the independent variable around the
regression line
Significance tests: test t, test F, Durbin-Watson, etc.
66. 66
Linear regression: cautions and limits
Three dangers:
1. Existence of self-correlation between
independent variables
2. Existence of hidden variables
3. Correlation does not imply causation
67. 67
Example (1/2)
Ice cream shop
1. Daily demand and
temperature are known
2. Hypothesis: there is some
correlation between
demand and temperature
Date Temperature (°C) Ice cream demand
68. 68
Example (2/2)
Estimate of coefficients
Variable Coeff Std Error T-Stat Signif
**********************************************************
1. Constant term -351.403 57.783 -6.08 0.000
2. TEMPERATURE 22.032 2.104 10.47 0.000
Significance indicators
Centered R**2 0.7965
Adj-R^2 0.7865
Regression F(1,28) 109.5690
Significance Level of F 0.0000
Durbin-Watson 2.1537
The estimates of
coefficients and the
regression indicators
suggest that the model
explains the
phenomenon well.
0
50
100
150
200
250
300
350
400
450
500
15 20 25 30 35
Icecreams
Temperature (°C)
69. 6969
Market tests
Experimental design
Testing with customers
Observation and registration
In case, what-if analysis
Demand forecasting