Document not to be Quoted/Cited

XIMB 2008-10 Batch

Jaccard analysis

For a 2x2 table with 4 cell frequencies a, b, c, d Jaccard = a/ (a+b+c), it is

possible to generate a similarity measure for pairs of heath facilities type and
for pairs of attributes (http://www.marketresearchworld.net). "The Jaccard
Analysis establishes which attributes are important to consumers and the extent
to which each of the channels is managing to satisfy these expectations"
(Chandrasekharan & Rajagopalan, 1989; Embretson, 2006, Streamline Surveys
Inc, 2008). Jaccard analysis results into Jaccard Similarity Coefficients which
are also referred as Jaccard scores. Jaccard analysis is done to find out the
variables in order of their importance for their ability to govern groupability or
similarity with the dependent variable or numerator a, as put in the above
formula. In this study, Jaccard scores were calculated as following:
A YY score = those who prefer at overall level or consume the brand and
associate the brand with a particular attribute
A YN score = those who prefer at overall level or consume the brand but do
not associate the brand with a particular attribute; and
A NY score = those who do not consume or prefer the brand at overall level
but still associate the brand with a particular attribute
Following equation provides the jaccard scores, also known as jaccards.
Jaccards = YY / (YY+YN+NY)
In Jaccard analysis, the higher the Jaccard score, the more important is the
attribute.

36

Document not to be Quoted/Cited

XIMB 2008-10 Batch

SWOT analysis

To understand the relative strengths and weaknesses of elements/attributes,

SWOT (strengths, weaknesses, opportunities and threat) analysis is done
(Winas, 2008; Crow et al, 2008). SWOT scores are calculated as following:
SWOT Score for a Brand
on Attribute A = ( Score of attribute A Attributes average ) (Average of Brands scores on all attributes
being considered-Grand Average of scores)
Attributes average is average of all brands scores for that particular attribute
(Row Average).
Average of Brands scores on all attributes being considered= Brand Column
Average
Grand Average of scores= Average of averages (can be calculated both wayseither as average of row averages or as an average of column averages)
Therefore, SWOT Score for a Brand
on Attribute A= Relative Attribute distance-Relative Overall Distance

In SWOT analysis, strengths and weaknesses are internal to organizations. For

strengths and weaknesses, scores are calculated. A positive SWOT score
calculated on the basis of abovementioned formula indicates strengths while
negative represents the weaknesses. The SWOT works on the principle of Zero
Sum game. If one organization or brand gains on an attribute, competition will
have score in negative to counterbalance that. Opportunities and threats are
external to organizations or sectors and inferred.

37

Document not to be Quoted/Cited

XIMB 2008-10 Batch

Stochastic Share Analysis

Step 1: Get the preferred set size for each brand preferers- Average
preferred set size
(This means to know e.g. how many brands are being preferred on an
average among Coca Cola preferers, among Pepsi preferers and so on) Say
X1, X2,, Xn
Step 2: Get the size of each brand preferers- e.g. proportion of coca cola
preferers in the sample, proportion of Pepsi preferers and so on. Say Y1,
Y2, Yn
Step 3: Divide size of the preferers by preferred set size for respective
brands and get the ratio for each brand
Y1 / X1= Ratio for Coca Cola =R1
Y2 / X2= Ratio for Pepsi=R2
And so on.
Step 4: Summate all the ratios.
Sum R= R1 + R2 + ..+ Rn
Step 5: Get the stochastic share
For Coca Cola, (R1 / Sum R)*100= CCs share
For Pepsi, (R2 / Sum R)*100= Pepsis share
And so on

38