Copyright INSIDE STORY 2007
1
Case Study:
Choice modelling with decision support tool
Background and objectives
â In a market driven by low prices and
deals, our client was losing ground to
competitors'offers that customers
perceived as a better deal even though
these "deals"were not necessarily
cheaper for the customer.
â Our client sought to strengthen their
market share by restructuring their offers
to better compete in the market. Our client
needed to:
æ Better understand which pricing
structures were perceived as 'better
deals'among each of their key
market segments.
æ Optimise their portfolio of offers to
cover the breadth of the market, and
at the same time not have offers that
cannibalised one another.
â Objectives were to:
æ Determine which pricing constructs
should be developed for the market,
including which offers will compete
effectively against competitors'offers
æ Determine which offers will ensure
spread of preference for our client's
offers across all market segments
æ Determine how different pricing
constructs compete with one another
æ Provide price elasticitiesfor different
price constructs
æ Quantify likely take up at specific
price points
æ Model the impacts on market share
of different pricing scenarios -
including anticipated reactions from
competitors
æ Provide a basis for optimising our
client's portfolio of priced offers, so
as to optimise their market share
and profitability
â This conjoint project was very large in order
cover all major offers among competitors in
this market as well as the many potential
pricing structures being considered by our
client.
â Required output included a decision tool
(market model), using conjoint analysis
Project details and outcomes
â A small qualitative phase was followed by a
major quantitative study
âQualitative phase : Conducted to confirm
and clarify current behaviour drivers, to
identify nomenclature used in pricing
structures and to gain preliminary
evaluation of potential pricing constructs in
preparation for the quantitative study
æ Focus groups were conducted
covering the main segments
âMajor quantitative phase : Major study to
quantify likely take up, measure price
elasticitiesand build a model of the market
(decision tool)
â Online survey (virtually all in this market
were regular internet users).
æ Online survey was programmed in
house. All data analysis was
conducted in house
â Respondents randomly recruited with
quotas for representative demographic and
geographic spread. Data was post-weighted
to known market proportions.
Advantage of Best-Worst Scaling over
conventional paired or choice conjoint is the
amount of information gathered in each
question. ie: when respondent indicates their
most favoured and least favoured out of A,
B, C and D, then we have results for five of
the six possible paired comparisons.
Thus Best-Worst Scaling more than doubles
the amount of data gathered per question.
Copyright INSIDE STORY 2007
2
â We used choice based conjoint analysis,
using Best/Worst Scaling: ie:
Respondents were shown sets of priced
options. For each set they indicate the
best for them (most preferred) and the
worst.
â Client's proposed list of offers were
deconstructed into their component parts,
so that utilities for each variable (attribute)
and each price point (level) could be
calculated.
â Fractional factorial design reduced the
number of product offers required to cover
the full breadth of potential offers.
â Even so, the required list of offers
required for this conjoint design was large
-likely to cause serious respondent
fatigue, which would compromise the
data.
â Our solution to making the questionnaire
more manageable for respondents (hence
preserving respondent freshness) was
twofold:
æ First we split the sample into
matched cells -each sample cell
was exposed to only 12 sets of
choices thus reducing the number
of choices any one respondent
needed to deal with.
æ The 12 choice questions were
spaced in between other questions,
so respondents did not face more
than 4 choice questions in a row
without a mental break.
â Sample cell rematching was based on
respondents'gender, region, brand
preferences and choice behaviour. After
the sample cells were rematched, the
effective total sample was n=300.
â Conjoint analysis provided value scores
(utilities) for each level of each attribute of
a product.
â Output was an interactive model of the
market using multinomial probitconjoint
preference simulation.
â This interactive model showed the likely
market share results of alternative pricing
scenarios, plus penetration of individually
priced offers (both our client's offers and
competitors'offers). (Example overleaf).
Conclusions and recommendations
â Output identified priced offers which were
able to compete more effectively with
competitors'offers, versus which offers
would simply transfer customers from
within the client's own customer base
â Output clearly demonstrated the impacts
of price level as well as price structure on
likely market take up.
â The decision tool allowed our client to
model impacts of possible moves from
competitors. This allowed them to identify
which priced offers would best protect
their own market share against potential
competitor reactionary discounting.
â Client was able to include the profitability
of each offer/price point and thus identify
which combination of offers would
optimise their profitability and market
share
â Enabled the client to identify which
combination of priced offers to launch in
the market
Case Study:
Choice modelling with decision support tool
Copyright INSIDE STORY 2007
3
Case Study:
Choice modelling with decision support tool
â Example of interactive market model:
Ability to insert or remove
alternative offers for each brand
Ability to set prices or size
of bonus/discount, etc
View results in total market
or in sub-groups/segments
View preferred choices
View brand market shares
View market share of offers
Client Results
â As a direct result of this conjoint study, our client launched (and have sustained) a
balanced portfolio of 5 enticing price structures
æ Their market position has now restrengthened and at the same time gave these
offers have given them a marked increase in revenue.
æ Our client no longer needs to run discount promotions, as they are able to attract
and hold customers based on their standard array of enticing offers.
æ They are no longer losing customers to competitors'deals, despite increasingly
aggressive price offers from some of their competitors