1、CONFIDENTIALMarket Research ToolkitMARKETING SCIENCE CENTERSeptember 1995ST-ZXe247/950830SeaHR1TABLE OF CONTENTSQualitative Focus groupsIn-depth interviewsObservational researchQuantitative Quick hitBasic A&UConcept testAttitudinalImage engineeringConjoint analysisDiscrete choiceCustomer serviceDiar
2、y panelST-ZXe247/950830SeaHR12MARKETING SCIENCE CENTERJohn Forsyth STMike Sherman NYKevin Nuffer NYAnil Kaul NYMark Leiter NYKK Davey NYJoshua Goff NYRich Milliman ATNeil Allison SFLinda Middleton CHSudeep Haldar CHXavier Azalbert LNAngela Lovejoy LNSid Simmons LNNirmala Santiapillai LNBernhard von
3、Skerst HAMark Esser DUPeter Prins AMRama Bijapurkar BYThe objective of this document is to provide a quick overview of marketresearch techniques. More detail can be found in the Market ResearchHandbook (PD Net #6229) or by contacting a marketing science specialist.ST-ZXe247/950830SeaHR13TWO TYPES OF
4、 TECHNIQUESQualitative QuantitativeObjective Generates hypotheses on decisionmakers, buying process, and keybuying function (KBF)Tests and quantifies hypotheses Tools available Focus groupsIn-depth interviewsObservational researchQuick hitBasic A&UConcept testAttitudinalImage engineeringConjoint ana
5、lysisDiscrete choiceCustomer serviceDiary panelThere are two types of market research techniques: qualitative andquantitative. Generally, a research program starts with qualitative to generatehypotheses and quantitative to prove them. ST-ZXe247/950830SeaHR14What is it? Approximately a two hour discu
6、ssion among 8 to 12 people, led by amoderatorWhen do you use it? Early in the research process, to clarify the issues, generate ideas onhypotheses, and help determine what other research to conduct. This isparticularly the case when there is little information on the subjector, when time and money a
7、re very limited, as the only research tool particularly when you dont expect a wide divergence of viewsHow do you do it? Hire a moderatorRecruit participantsPrepare a discussion guideAttend focus group(s)Videotape the groupsDebrief moderator to determine conclusionsWhat do you get out it? A summary
8、of key findings and well developed hypotheses for answeringcertain key issuesApproximate cost $3,000 - $5,000 for each group of 10 consumer respondents Project typically two to six groupsApproximate time To develop discussion guide, generate hypotheses: 2 weeksTo conduct: in 1 week you can expect to
9、 conduct three to four groups if several cities are involvedTo synthesize findings/refine hypotheses: 1 to 2 weeksFOCUS GROUPSST-ZXe247/950830SeaHR15FOCUS GROUP OUTPUT PAGEIssue Discussion output Quotes ImplicationsSales force Customer think salespeopleare lazyMethods to improve thesales force did n
10、ot surfaceThey never follow upIm not even sure whomy sales guy isNeed to focus thesalesperson portion of thequestionnaire on how toimprove the sales forceCompetition Little agreement as to whothe competition isWe compete with XYZhead to headI never lose sales toXYZ. Dillon and Petersare my main comp
11、etitionLet respondent define theircompetition on questionnaire do not define it for themHave section in questionnaireto find out who competes withwho, and whyST-ZXe247/950830SeaHR16What is it? One-on-one interviews usually lasting at least 45 minutesPrimary research tool when only have a few custome
12、rsWhen do you use it? Similar to focus groups, in-depth interviews are used to: Explore issues and underlying details Develop hypotheses Gain an understanding of key buying factors and brands Discover the language customers useLikewise, they can add value after quantitative research to further get y
13、ourhands around the findingsOne-on-one are preferred to focus groups when: There exists specific people you wish to speak with Qualified respondents are few in number Specific issues requiring confidentiality, sensitivity or probingHow do you do it? Hire a moderatorRecruit participantsPrepare a disc
14、ussion guideDebrief moderatorWhat do you get out it? A summary of key findingsWell developed hypothesesPotential product/service definitions for quantitative researchApproximate cost Projects typically $100 to $300 per interview Approximate time Interviews usually last 45 minutes to 2 hoursIN-DEPTH
15、INTERVIEWSST-ZXe247/950830SeaHR17INTERVIEW WITH JOE BLOGGS, 34 YEARS OLD, 2 KIDS, C2On February 1, 1994 we met with Joe Bloggs and his wife. The purpose of thediscussion was to understand the decision making process they go through whenbuying financial services. Key points1. They make a joint decisi
16、on when buying insurance. 2. They like the appeal of a one stop shop with one company. 3. They dislike home service and find it uncomfortable but instead prefer using the telephone. Issues1. An in-depth study of cross-selling opportunities and appeal would be beneficial. 2. A more thorough understan
17、ding of selling channels would be useful. Next stepsQuantify segment size and scale of opportunitiesST-ZXe247/950830SeaHR18OBSERVATIONAL RESEARCHWhat is it? A qualitative technique to observe customersWhen do you use it? To identify potential new products and servicesWhat do you get out of it? Hypot
18、heses on latent demand from observing behaviorand problems customers faceApproximate costs VariesApproximate timing VariesST-ZXe247/950830SeaHR19Many companies are observing decision makers for new insights into thecustomer. EXAMPLES OF OBSERVATIONAL RESEARCH A pharmaceutical company followed GPs ar
19、ound all day to identify problems/stresses they face. Honda sent R&D engineers to rent rooms in houses in California to see how Americans interact with their cars. Small video cameras in retail stores record how consumers make decisions at the shelf. Ivory soap was created by seeing women who tended
20、 to lose the soap when washing clothes in the Ohio River. ST-ZXe247/950830SeaHR110Listed below are some best practices for using qualitative research. HELPFUL HINTS FOR QUALITATIVE Generate hypotheses before qualitative to help structure research. Take stimuli (e.g., pictures) into qualitative to fa
21、cilitate discussion. Be sure to be complete. You need to have an exhaustive list of key hypotheses. Attend the sessions. Videotape them for others not able to attend. Develop a first-hand point of view. Do not be too literal. People frequently are unable to articulate their precise thoughts. ST-ZXe2
22、47/950830SeaHR111Quantitative studies seek to prove hypotheses with data from a statisticallyrepresentative sample. QUANTITATIVE TECHNIQUES Quick hit research Is a short, very focused quantitative surveyBasic quantitative(usage and attitude)Measures basic beliefs and behaviors for a categoryConcept
23、testing Judges consumer reactions to specific new product ideaAttitudinal Measures category attitudes for latent demand situationsImage modeling Analyzes effects of brand imageryConjoint analysis Measures trade-offs of product/services features and priceDiscrete choice Measures in-depth trade-offs o
24、f a few features and priceProblem detection Understands root causes of why customers leaveCustomer servicemodelingPrioritizes specific elemenst of customer serviceDiary panel Track consumer purchases over timeST-ZXe247/950830SeaHR112QUICK HIT RESEARCHWhat is it? A very short quantitative questionnai
25、re done veryquicklyWhen do you use it? To answer a very focused set of issuesWhat do you get out of it? A quantitative answer quicklyApproximate costs Typically under $20,000Approximate timing Typically 1 to 3 weeksST-ZXe247/950830SeaHR113EXAMPLES OF QUICK HIT RESEARCH Interviewing at conventions to
26、 target a hard-to-reach group (e.g., ConventionResearch Studies) Adding a few selected questions to omnibus studies (e.g., Telenation atMarket Facts) Targeting hard-to-reach consumer groups in a touch tone telephone survey(e.g., Yankelovich Project Advantage) Identifying new products and services op
27、portunities among a key customergroup, e.g., club member Profiling corporate identities among womens issues Determining price awareness for an impulse productMany companies use quick hit research for a very focused set of issues. ST-ZXe247/950830SeaHR114Turning to Yankelovich as a source, we have a
28、strong relationship (in someways, partnership) with Yankelovich to conduct quick surveys in theircontinual phone interviews drawing on a 16,000 respondent database (seeseparate document, McKinsey & Yankelovich Partnership on Project Advantage by Kevin Nuffer, January 1995). OVERVIEW WHAT IS IT?A res
29、pondent database . . . linked to . . . a rapid turnaround surveycapability 16,000 households Extensive background data Demographics Econographics Product usage Media habits Attitudinal battery Capability to conduct surveys up to 10 minutes long viainteractive voice response (IVR)telephone system Int
30、elligent sampling from database Surveys completed in 2 or 3 days start to finishST-ZXe247/950830SeaHR115What is it? Survey that measures basic beliefs about a category and identifies groupsof customers by their needsWhen do you use it? When purchase decision is straightforwardWhen client has never d
31、one research before and has little quantitativeinformationHow do you do it? Hypothesize key buying factorsDevelop a range of KBF statementsMeasure consumer importance and competitive brand positioning on theKBFsIdentify segments by KBFsLink segments to reported purchase behaviorCross reference with
32、other data to identify gapsWhat do you get out it? A better understanding of customers perceptions of products/services inthe marketplace and some ideas for ways to position productsApproximate cost $30 to $150 per consumer respondentsApproximate time 1 to 2 weeks for qualitative research6 to 10 wee
33、ks to design, execute, and analyze researchDESCRIPTION OF TOOLS BASIC QUANTITATIVE (USAGE AND ATTITUDE)ST-ZXe247/950830SeaHR116TYPICAL QUESTIONSBuying factor Importance* Brand A Brand B Brand CGood serviceLowest pricesBroad selection ofstylesKnown brandsLatest fashionsCompetitive ratings on delivery
34、 of the buyer factor*The usual A&U questions probe the importance of key buying factors andcompetitive perceptions on each. * Scale of 1-10ST-ZXe247/950830SeaHR117COMPETITIVE RATINGS FOR A RETAILER* 10 point scaleILLUSTRATIVEGood serviceLatest fashionsBroad selection of stylesLowest pricesKnown bran
35、ds6.46.46.05.65.5KBF Importance* 0 1 2 3 4 5 6 7Rating*All have weakserviceBrand strengthis fashionability,selectionWeakness ispriceRange of competitionOur brandFor example, this retailer has a competitive strength in fashionability but isweak on price. ST-ZXe247/950830SeaHR118CASUAL WEAR NEEDSAvera
36、geSource: Market research-0.4-1.60.22.10-10-0.601.702.9-0.2-1.4-0.20.80.60.2-0.9-0.6Good service 6.4Latest fashions 6.4Broad selectionof styles6.0Lowest prices 5.6Known brands 5.5ILLUSTRATIVETotalaverageSegment A Segment B Segment C Segment DHowever, results should be understood by segment, each of
37、whom have adifferent priority of needs. ST-ZXe247/950830SeaHR119CONCEPT TESTINGWhat is it? A research technique based on one-on-one interviews whereconsumers react to a specific new product ideaWhen do you use it? When you need to assess a specific new product ideaWhen you are confident you have def
38、ined the new idea properlyHow do you do it? Write a concept (similar to a print ad)Conduct interviews: Current brands used Habits and practices Read concept Intent to purchase Likes and dislikes Image ratingsProject share/volume with calibration factorsWhat do you get out of it? Share/volume for new
39、 product ideaApproximate cost $75 to $300 per interview Approximate time 6-10 weeks to design, execute, and analyze researchST-ZXe247/950830SeaHR120CONCEPT TESTINGPictureDescription of the new productidea including its benefits,price, brand. This should besimilar to a magazine ad butwith less advert
40、ising hypeQuestions asked Purchase intent Likes and dislikes Hedonic Price/value Uniqueness Preference versus potential competition Perceptions on KBFs Image ratingsIn a concept test, respondents reactions to a new idea are gauged. ST-ZXe247/950830SeaHR121CONCEPTS CAN BE CALIBRATED TO TRIALPurchase
41、intentPercent ofrespondents Calibration*Definitely would buy 50% 75Probably would buy 30 35Might or might not buy 15 5Probably would not buy 5 0Definitely would not buy 0 0* Some research agencies have good experimental databases on calibration. Numbers are culturally and sometimes category dependen
42、t= 48.8% 19.5%x 50%x 80% Interested universe Awareness Distribution Predicted trialConcept purchase intent can be converted in trial rates, though the calibrationfactors are a key assumption. Certain market research firms specialized in thisand have good databases and experiences. A marketing scienc
43、e specialist canhelp you find the appropriate firm. ST-ZXe247/950830SeaHR122STRETCHING CAN BE EVALUATEDCurrentbrandNewconceptKBFsAble to stretch thesekey elementsBoth have sameweaknessesBy evaluating the perceptions of a new concept versus a current brand, we cananalyze when it is possible to stretc
44、h a brand and where the brand is too strongto move. Gives good serviceIs prestigiousHas latest fashionsHas lowest priceHas known brands0 1 2 3 4 5 6 7Rating23ATTITUDAL RESEARCHWhat is it? Analysis based on individual interviews that identify groups of customers by theirattitudes and perceived benfit
45、s toward the product or service or the productcategoryWhen do you use it? In latent demand situations, when a market does not currently exist for the productWhen decisions involve complex issues such as multiple categories and futuretrendsHow do you do it? Hypothesize key attitudes and develop state
46、mentsMeasure customer agreementIdentify segments by attitude needsLink segments to reported purchase behaviorCross reference with other data to identify gapsDetermine likely acceptance of new/revised value propositionsWhat do you get out of it?A better understanding of customers attitudes toward pro
47、ducts/services in themarketplace and some ideas for ways to leverage those attitudes to positionproductsApproximate cost $50 to $300 per consumer interviewApproximate time In one week 300 half hour phone interviews or 200 half-hour personal interviews1-2 weeks for qualitative research6-10 weeks to d
48、esign execute and analyse researchST-ZXe247/950830SeaHR124Attitudinal segmentation is based on agreement to attitude/need statements. Each statement expresses customers agreement towards a specific idea. EXAMPLES OF ATTITUDE STATEMENTS1. Telecom services are essential for my business2. It is very im
49、portant for my business to receive bills for each extension3. It is not necessary to understand my business to sell me telecom services4. Mobile phones are very critical for proper functioning of my business5. We definitely prefer to buy all our telecom services from one vendor6. It is very importan
50、t for my business to have direct numbers for employees7. We prefer to deal with a telecom company that has good service rather than just a low price8. If my employees are traveling they should be able to be in phone contact with their home office9. I would definitely prefer to have my telecom compan
