Please enable JavaScript.
Coggle requires JavaScript to display documents.
SECONDARY & PRIMARY DATA (I) (Sources of Secondary Data (External …
SECONDARY & PRIMARY DATA (I)
Sources of Secondary Data
Internal
Co. records
Sources of Internal Secondary Data
Sales
--> Aggregate (across custs by sales period, how sales period have changed by month/ yr)
--> For indiv custs
Prices
Advertising (are we getting more sales w more ads?)
--> Expenditures
--> Creative used
Cust Databases
--> Feedback (ability to adjust and respond to the feedback given, is that the org's prob?)
--> Purchase history
What research qns could be answered using this data?
Cust characteristics that will increase the probability of engaging them in purchasing X pdt
Would attractive ads drive sales? Who are the custs who are responsive to the ads shown?
Appeals --> test whether a fear appeal speaks more to the custs compared to another emotion
Forecasting sales --> how much of a particular item we should have in stock --> prevents stockout when custs look for pdt --> deters custs from coming back and repurchasing from you
External
Publicly available
Commercial
External Secondary Data: Public Data
Public Data
: info that can be freely used, reused and redistributed by anyone w no existing local, national or international legal restrictions on access or usage
E.g.
--> Internet
--> Govt publications
--> Publications by foundations, trade associations, unions
External Secondary Data: Commercial
Retail store audits
--> send team to store to monitor
-sales
-distribution
-prices
-support
Consumer purchase panels
--> home audit
--> mail diary audit (ask custs to keep receipts from what they have purchased)
Scanner data
(barcodes)
--> track purchases in (large) no. of stores
--> sales, prices, coupons etc automatically recorded
Single-source data
(tracking movement across webpage, better unds custs journey on online stores)
--> track panel members' purchases and media consumption over time
[Commercial data collected by co. and sold for profit e.g. Nielsen]
[Time consuming]
[Focus on the fact that who is paying them to get data]
[Timeliness and cost]
Why Use Secondary Data?
save $
save time
proven data collection method and quality
may be only source available e.g. clothing brand introduced on Amazon, what other brands are they looking at?
Can You Trust Secondary Data?
Data quality can be influenced by:
Data collection method
e.g. small sample size used by IS Bureau of Labour Statistics
Purpose of data collection
Source of data collection
e.g. who collects the data may influence the way in which the DV is operationalised
-->
VIMPT!!! Always ask yourself: "WHY and HOW were the data collected?"
-maybe collected to please the client
-which state and what time of day did they collect the data --> influence the quality of the data
-ppl may not be honest when asked abt the cigarettes --> sales of cigarettes reflects a more true data
Primary Data (1): Qualitative Data
Qualitative Data
Data that approximates and characterises
Typically:
--> non-numerical in nature
--> obtained thru observation, interviews, focus grps etc
--> small no, of non-representative cases
--> unstructured
--> analysed using non-statistical methods
--> used to develop an initial understanding
Acceptable uses:
--> exploration e.g. identifying and defining the research problem
--> orientation i.e. understanding the cust's perspective (how they see and talk abt it)
--> gaining insights into topics that are difficult to pursue w structured research
Observation
Example methods:
Spying (done ethically)
e.g. an employee observing patrons in starbucks
e.g. CCTV can be used to determine if the elderly take longer at ATMs
Mystery shoppers
--> hired by the co. to check up on the service of the employees
--> the no. of items you sell to the custs (pair earrings w necklaces)
--> give feedback on how employees performed (competition --> extra pay, extra day-off)
Possible insights:
Elderly ppl find ATMs difficult to use
--> step by step guides pasted on ATMs
--> use high contrasting colours
--> enlarge font
Many custs read the pdt labels
--> supermarkets can enlarge the nutrition labels that custs look at the most
--> highlight certain attributes you want custs to see easily, see if pdt fits custs' preference --> healthy pdts, common allergies (nuts, eggs), halal, vegan
Kids decide which candy to buy
--> request for placement at eye level of kid (lower shelves)
--> sample for kids to try the sweets
--> diff pricing strategies for premium pdts
Hotel receptionists look at their screen instead of guests
Observation
Learn sth by watching your custs and employees w/o intruding their behaviour
Adv:
--> not influenced by socially desirable responding (if ppl want to be seen in a +ve light, will say e.g. they smoke 10 instead of 40 cigarettes a day)
--> least expensive and most accurate method for collecting behavioural data
--> sometimes the only viable method
e.g. when we need to unds the behaviour of v young children (unable to verbalise their preference)
Ethnography
Ethnography: scientific description of ppls and cultures w their customs, habits, and mutual differences
From Anthropology
--> study of humans and human behaviour and societies in the past and present
--> researcher immerse themselves in cust experience
--> done by trained anthropologists or sociologists
Objectives
--> direct observation of cust
--> experience ins and outs of custs' daily lives
Harley-Davidson
Employees join custs on ride alongs and gain deeper insight on how custs feel abt their bikes
When they stop at restaurants, they sit near the window to keep their eyes on their bikes (constantly looking at their bikes, don't want it to be damage)
At end of meal, they go up to service counter and ask to wipe their bikes, keep bikes shiny and clean
HD used these insights to maintain their bike services, improve their services
--> Knock down walls b/w waiting area and service area so custs can see how well their bikes are being treated
--> Free cleaning of bikes
--> Service ratings increased dramatically
In-depth Interviews
In-depth engagement w indiv to unds his/her views
2 types:
-->
Non-directive
: the respondent is given freedom to respond within the bounds of the topic of interest
-->
Semi-structured or focused
: the interviewer attempts to cover a specific list of topics or subareas and the respondent has less freedom to extend the bounds of the topic
Adv:
Givs each participant equal "air time"
Appropriate for sensitive topics e.g. sexual behaviour
Appropriate for topics where sub-culture may be an issue e.g. strong social norms/ peer pressure
Comprehensive, get a lot of info of what participant is thinking
e.g. fast food chain --> should introduce more vegetarian options: focused grp will be downed out by the rest bc of the -ve social stigma attached to these preferences, so these ppl may feel threatened to speak up. In-depth interview allows them to open up and you can then better service their need.
Disadv
:
Difficult to get relevant info
Time consuming
--> may go off track and talk abt sth they are passionate abt
--> busy interviewees, esp executives so need to be specific
Lacks generalisability
--> bc small grp, cannot generalise beyond sample
Heavily dependent on interviewer skill
--> remain calm, patient and keep them on topic, don't cut them off
--> Avoid steering the convo (too much)
--> Be objective and neutral
Participant motivation e.g. maximising own (future) utility
--> best quality for cheapest price, ppl have other motivations at play that will affect their response, may be dishonest bc they want a particular outcome
Focus grps
Obtain ideas/ solutions from a small grp thru discussion
--> participants (5 and above) are encouraged to express their views and elaborate on views of others
Moderator Role
--> Passive participant (relative to interviews, don't be actively involved in the discussions)
--> Encourage participation (encourage everyone to speak up and contribute equally in the discussion)
--> Prevent social influence (can change the way ppl are expressing their opinions)
Adv
:
Richness of data
Versatility
Encourages participants to speak-out
Impact on managers
Usefulness of data
Ppl feel more comfortable in these situations to speak out, works well here but not on sensitive issues
Disadv
:
Lacks generalisability
Potentially misleading data
Social influence impacts responses
Participants have other motivations e.g. maximising own (future) utility
e.g. Pepsi edge (50% sugar) --> ppl didn't unds positioning of pdt so custs didn't purchase and it was a failure
Focus Grp Challenges: Conformity
Conformity
: a type of social influence involving a change in belief or behaviour in order to fit in w a grp
Why do ppl conform?
--> Normative conformity
-Desire to fit in/ fear of rejection
-Public
-Competitive adv in the real world, someone who has the resources likes you, you may get access to the resource
--> Informational conformity
-Ppl lack knowledge
-Private and public
-Want to have correct info that we are communicating to other individuals --> affects behaviour in public and private
Factors Affecting Conformity
Conformity more likely when
:
The judgment is difficult and public
Ppl are uncertain and have low status (might increase in likelihood of being accepted by grp)
Grp size is large
Rest of grp is unanimous
--> when we feel that the grp has better insight than us and we are concerned of being wrong
Conformity less likely when
:
The judgment is easy and private
Ppl feel confident and have high status
Grp size is small
Rest of grp is disagreeing
Strategies for Preventing Conformity
:
Keep qns simple and to the point
--> judgment is easier so less likely for ppl to conform (difficult qns where ppl dk opinion will increase conformity)
Ask for personal experience/ opinion
Reverse response order
Be neutral
--> don't give indication of who low status person is
Have supportive body language
Small grp so voices don't get drowned out
Encourage dissenting opinions
Focus Grp Challenges: Obedience
Obedience
: a form of social influence in which a person yields to an authority figure
Why do ppl obey?
--> Norm not exception
--> Ppl even forgo financial gain to obey authority
--> sth we are socialised to do, been taught this since a young age, helps in the society
Milgram's Obedience Experiment
Factors Affecting Obedience
Obedience more likely when:
The authority figure is legitimate
--> expert knowledge
--> higher in social hierarchy
Participant is high in authoritarian submission
Participant is low in social intelligence
Obedience less likely when:
The authority figure is legitimate
--> no expert knowledge
--> lower in social hierarchy
Participant is low in authoritarian submission
Participant is high in social intelligence
Strategies for Preventing Obedience
Soft Tactics
Avoid being seen as the expert
Reinforce the fact that the respondent is the expert
--> put the expertise in the hands of the participants by telling them they are the experts in this area
--> ensure you are one of the participants and not coming across as the leader of the grp
Highlight respondent unique knowledge
Hard Tactics
Reduce hierarchical distance
Dress similar to participants
Sit among participants
Participant Selection
Select participants based on personality
Don't select participants high in authoritarian submission
Don't select participants low in social intelligence
Analysing Qualitative Data
Interpretation (Level 1)
--> What do respondents say/ feel/ do?
Analysis
--> Sitting, differentiating, separating, ordering the data
Interpretation (Level 2)
--> What patterns ermerge? What do they mean?
Interpretation (Level 3)
--> What does it mean in relation to client issues?
-Analysis is subjective and influenced by researcher expectations.
-Follow-up by testing hypotheses using quantitative data.
-Opinions categorised based on themes across the interviews and grp them tgt then sub categorise
-Subjective --> stand out themes are the ones researcher believes are the most impt, be careful when looking at data and interpretation of it --> have to be backed up by qualitative data
Pros and Cons
Pros
:
Deep insights
Useful for generating hypotheses
Cons
:
Small sample size
--> Representative?
--> Generalisable?
Significance testing not possible