L10, Data analysis 1 - quantitative Flashcards
Explain the differences between exploratory and confirmatory research in the early phases of product development
Explanatory research
— Understand why cause/effect relationships.
Confirmatory research
— Researchers has a theory - Hypothesis and the objective of the research is to find out if their hypothesis is correct
Process of quantitative data analysis
- create a data set, in an excel worksheet or in a dedicated statistical analysis software
- Clean up the data set
- analyze basic metrics and diagrams
- analyze relationship between two variables
- analyze relationships between three or more variables
- compare with data from interviews
Analysis of quantitative study?
- computation of standardized metrics
- visualisation
- interpretation
—> what does the number mean?
—> what can we say about the views of non-respondents - What conclusions can be drawn?
Analysis of correlations and open-ended questions are essential to explain reasons for ratings and suggest improvements
Analysis of quantitative study?
- computation of standardized metrics
- visualisation
- interpretation
—> what does the number mean?
—> what can we say about the views of non-respondents - What conclusions can be drawn?
Analysis of correlations and open-ended questions are essential to explain reasons for ratings and suggest improvements
Name some basics metrics
1. Measures of central tendencies —> mean —> median —> mode —> sum —> N —> Response rate
2. Measure of variability —> range —> inter-quartile range —> variance —> standard deviation —> standard error —> min —> max
Name some basics metrics
1. Measures of central tendencies —> mean —> median —> mode —> sum —> N —> Response rate
2. Measure of variability —> range —> inter-quartile range —> variance —> standard deviation —> standard error —> min —> max
name some types of analysis of relations between two variables!
- Tabulating and cross-tabulating proportions
—> e.g. agreements with statements for owners of different brands - comparing means across items, groups of customers
- correlation coefficients
—> predicting an outcome as a function of an antecedent variable
—> e.g. level of satisfaction as a function of length of relationship with vendor
What does survey monkey provide?
- enables cross-tabulations
- selection
- evaluation of statistical significance
- text analysis
Strengths and weaknesses with quantitative data analysis?
STRENGTH
- provides an investigation with scientific status
- Offers confidence in the findings
- precise measurements
- enables analysis of large volumes of data
- provides concise presentation of data
WEAKNESS
- quality of data may be poor even if quantified
- you may be overloaded by data
- quantitative analyses is not as scientific as it might seem on the surface
Why should you do chice modeling/ conjoint analysis, concept testing or experimentation in product planing?
- what drives the choice of one product configuration over another?
- What are appropriate attribute levels?
- How much are they willing to pay?
- How many would buy at the price?
Give some examples of choice modeling tasks!
EXAMPLE 1: Choose TV Attributes: 1. type — plasma —> LCD —> LED
- size
—> 36””
—>40””
—> 46”” - brand
—> sony
—> toshiba
—> philips - price
—> 499
—> 699
—> 899
Attributes:
- type
- size
- brand
- price
Product profiles:
Then different levels for each attribute
EXAMPLE 2: glases Attribute: 1. Lens type —> polarising —> UV protector —> prescription
- Design
- Price
- Frame type
- Lens color
—> brown
—> blue
—> yellow
—> black
6. Brand —> rayban —> Oakley —> D&G etc.
Describe the conjoint analysis procedure
- Identify attributes of the product
- Decide on how many levels that will be considered for each category
- Create screen shots or cards fr each variant you want to examine
- Determine judgement procedure
—> pairwise comparison?
—> preferential scale?
—> probability to purchase - Administr survey
- Compute utility weights for levels of attributes and attribute importance for individual responses
- Aggregate responses
—> in market segments?
—> cluster analysis
what can be done to minimize the number of permutations?
fractional factorial design of experiments can be used
Important to think about when administering a survey?
How?
—> web
—> e-mail?
—> postal?
Collect addiational information:
- spending level
- involvement with product
- purchae plans
- demographics
How do you calculate attribute range?
max utility-min utility
How do you calculate attribute importance?
attribute range/ sum range
Strengths and weaknesses of choose modeling?
STRENGTH
1. Can handle complex relations
—> which can be hard to estimate otherwise
- Good beyond a customer’s self-report by forcing the respondent to act
- large number of product design alternatives
- mix of confirmatory and explanatory research
WEAKNESS
1. limited scope
—> sample size
—> choice of attributes
- time-consuming & costly
- does not fit all buying situations
—> e.g. decisions made by groups - modellable product complexity is limited to
—> about 6 attributes
—> 2-3 levels
How can you expand conjoint analysis into controlled
- compare alternative price points or product designs
- Establish treatment groups exposed to different alternatives
- An alternative to conjoint analysis
- more predictive than surveys and focus groups
- fewer differences than in conjoint analysis
Concept testing?
- Testing of concept by soliciting a response to a description of the product concept from potential customers in the target market
- Face-to-face sessions with open-ended interactive formats are suitable for early phases
In what phase of the product development is qualitative concept testing appropriate?
- In the planning and concept development phase
In what phase of the product development is quantitative concept testing appropriate?
In the testing and refinement phase
Namn some purposes for concept testing!
- which of several alternative concepts should be pursued?
- How can the concept be improved to better meet customer needs
- approximately how many units are likely to be sold?
- Should development be continued?
Describe the concept testing process
- define the purpose of the test
- choose a survey population
- choose a survey format
- Select way to communicate the concept
- Develop survey format
- measure customer response
- interpret the results
- reflect on the results and the process
give example of concept testing set-up for electric scooter
- Purpose of concept test
—> what market to be in - Survey population
—> college students who live 1-3 miles from campus
—> factory transportation - Survey format
—> face to face interviews
Name some different ways to communicate the concept
- verbal descriptin
- sketch
- photos and renderings
- storyboard
- video
- simulation
- interactive multimedia
- Physical appearance models
- Working prototypes
Describe the steps of designing the survey format
PART 1: Demographics and behavior
- How far do you live from campus
- How do you currently get to campus from home
- How do you currently get around campus
PART 2: Product description
—> Present the concept description
PART 3: purchase intent
—> id the product were prices according to your expectations, how likely would you be to purchase the scooter within the next year?
—> definitely not, probably not, might or might not, probably will, definitely will
PART 4: Comments
- What would you expect the price of the scooter to be?
- What concerns do you have about the product concept
- Can you make any suggestions for improving the product concept?
How can you interpret the results from the concept testing?
E.g. estimation of sales volume
Q=NA(CdefFdef+CprobFprob)
check bulrush eppinger p.177-178
Name some conclusions for choice modelling
- choice modeling methods include conjoint analysis and experimentation
- conjoint analyses provides quantitative results and is mainly effective for optimizing an established concept
- choice modeling can be coupled to purchasing intent questions and then be used to estimate sales volume
