Fundamentals Of Market Research Techniques

2y ago
75 Views
4 Downloads
3.83 MB
33 Pages
Last View : 1m ago
Last Download : 2m ago
Upload by : Aiyana Dorn
Transcription

Fundamentals ofmarket researchtechniquesAn e-book fromCharterhouse Research

ForewordCharterhouse Research is pleased to bring you its Fundamentals ofmarket research techniques guide.The book is intended to be a basic step-by-step guide to market researchtechniques, designed for new-to-research client-side research personnel.We are often asked by our clients if we can talk through some keyprinciples with their research teams: for example, what makes a good brief?We hope you find our e-book a useful tool.Best wishes from the Charterhouse Research team.

A guide to market research techniquesAn e-book from Charterhouse ResearchBack toContentsContentsClick on each heading to jump straight to that page1. Briefing1.11.21.31.41.51.61.7The structure of an ideal research briefWhat are you trying to achieve from the research?What do you already know?Target research audiencePossible approach/methodologyWhat to ask?Other items to include/consider in the brief2. Main research methods available and samplingconsiderations2.12.22.32.42.52.62.7When can market research be of help? When ismarket research not appropriate?What methodologies should be used?Qualitative research - Why use it?Qualitative research - MethodsQuantitative research - MethodsTypes of quantitative study availableSampling considerations3. Quantitative questionnaire design andinterviewing3.1 Questionnaire design and ideal interview lengths3.2 Asking the right questions3.2.1 Golden rules3.3.1 Scales3.3.2 Nets3.3.3 Top box3.3.4 Mean scores4. Tables, weighting and statistical testing4.14.24.34.44.5How to check tablesHow to read tablesBase sizesWeightingSignificance testingCharterhouse Research StatsChecker5. Timelines for average market research projects6. Outputs

1 Briefing1.1 The structure of an ideal research briefBack toContentsThe ideal brief is broken down into the following sections:14710Introduction &backgroundTargetaudience5Topic coverage/what to archobjectivesPossible approach/methodologyDeliverables/output9Timescale

Briefing1.2 What are you trying to achieve from the research? Rubbish inRubbish out Business objectives:Research objectives: What is the ultimate aim to thebusiness of the overall project?For example, bring in 10,000new customers.Back toContents What are we hoping to learn from the research? What sort of questions are we hoping theresearch will answer? How do we intend to use the research results?

Briefing1.3 What do you already know?Back toContentsDo not reinvent the wheelNahhh, I don’t think it will work. Let’s do something different something smarter, something cooler!Before putting together aresearch brief, ask: What other information doesthe business have already thatanswers some of the objectives?What similar projects have youalready undertaken?What other non-researchinformation might answer thequestions?

Briefing1.4 Target research audienceWho? Who is the target audience? How are you defining them,e.g. qualifying criteria? Are there other stakeholders – i.e. staff, ex-customers,intermediaries – that need to be included? For business research, who within the target company will be thedecision maker? Are there any significant opinion formers/influencers?How do we find them? Does the internal client have his/her own source of sample/database of customers/prospects/intermediaries? Have they been developing target lists, etc? Does the research agency need to free-find? Do we have anyinternal data (e.g. penetration figures that will help the agencyfree-find)?Any sub-groups of interest? Do we want to be able to identify any differences bysub-group (e.g. life stage, location, gender)? Any must-haves versus nice-to-haves?Back toContents

Briefing1.5 Possible approach/methodologyNot all methodologies will be appropriate and logistics mightprevent you from doing what you ideally want to doObjectives Is exploratory research needed: will require qualitative method? Are robust numbers needed: will require quantitative research?Sample provision Does the sample allow for a quantitative methodology; is the universe big enough? Does the sample allow for clustering for f2f groups, or cost-effective f2f depths?Use of stimulus Do we need to show respondents something? Is a face-to-face methodology required – could we talk by phone/email/online?Timing What is the ultimate deadline for results? Do you need toplines, etc? Is it achievable with ‘the best’ methodology, or do we need to compromise?Cost What is the budget available? Is ‘the best’ methodology achievable within budget, or do we need to compromise?Back toContents

Briefing1.6 What to ask?Back toContentsThe respondent CANNOT directly answer your business objective for youThey CAN answer questions that are relevant to them inlanguage they can understand CompetitorintelligenceMarket researchINSIGHTANSWERDECISIONInternalinformation

Briefing1.7 Other items to include in the briefDeliverables/OutputTimescaleBudget outputs do you need? What Who is going to use the feedback? How are they going to use it?there any deadlines? Are Any timings of importance (e.g.communication mailing, product launch)is the ballpark budget? What Any compromises?Back toContents

2 Research methods available and samplingconsiderations2.1 When is market research appropriate and not appropriate?Back toContentsBack toContentsIn the majority of cases, market research can add value where a better understandingof the market, people’s motivations and attitudes can affect businessImportantly, primary research is not always appropriate There is insufficient budget There is insufficient time There is insufficient sample When results cannot be actioned When existing data is already availableWhen you can’t conduct the study effectivelybecause:Given strict rules set by the MRS, marketresearch is inappropriate as a: Lead generation exercise Direct sales/marketing tool

Research methods2.2 What methodologies should be used?Qualitative researchprovides the ‘why/howdo people think?’Quantitative researchprovides the ‘how manypeople think that way’ Relatively smallnumber of respondents Elicitation and Large number ofrespondents Allows for a better exploration of the fullrange of reactions,ideas, thoughts,perceptions the motivations,feelings andattitudes behindbehaviourNot statisticallyreliableunderstanding ofhow many customersdisplay the reactions/thoughts/ideas elicitedat the qualitative stage places these itemsin some sort oforder/hierarchyStatistically reliabledataBack toContents

Research methods2.3 Qualitative Research – Why use it?Back toContentsObjectives/purpose: To gain understanding of underlying reasons and motivations To provide insight into the background/context of a situation/problem, generating ideas/hypotheses to be tested inquantitative researchTo uncover trends in thought and opinion and help identifypotential market segmentationCan be conducted as a standalone research project or partof a mixed methodology Samples: Typically small numbers, not designed to be statisticallysignificant Samples are often structured with a minimum number ofrespondents in key quota groups (e.g. business size, industrysector, location, consumer socio demographics, etc.) rather thanstructured to be representative of the total marketData collection: Unstructured Semi structured Observation ParticipationQualitative research is used to answer theWhy? and How? questions

Research methods2.4 Qualitative Research - MethodsBack toContentsFocus Groups Cross fertilisation ofF2F depth interviews Allows for detailedTele-depth interviews Telephone widelyOnline forums Allows for crossviews and ideas: Ideal for product/service development Relies on availabilityof clustered sample ofrelatively homogenousrespondents Less suitable whereinformation is sensitive Not ideal forunderstandingindividual decisionmakingprobing and individualgranular feedback: Ideal forinvestigating whycustomers have madea particular decision In b2b environmentopportunity to observerespondent at work/consumer at home Interviews can bepaired/triads Can be costly andtime consuming tocomplete if respondentsare not geographicallyclusteredused method ofcommunication amongb2b audiences andthus tele-depths workwell Typically lower costthan f2f Interview lengthtypically shorter thanf2f interviews so lessdetailed feedback Difficult to establishthe same level ofrespondent rapportgained in f2f depthsfertilisation of viewsand ideas althoughlevel of detail gainedis less than f2f focusgroups Can be conductedover a number of days– ideal if some formof diary completionexercise is required Response rates toonline forums aretypically lower than f2ffocus groupsMystery/sensitisedshopping A form of participantobservation whichtypically usesresearchers (thoughsometimes researchrespondents, marketingstaff, businesscustomers, etc.) toact as a customer/prospect to monitor thequality of processesand procedures usedin the delivery of aservice

Research methods2.5 Quantitative Research - MethodsF2F Interviews Allows for stimulus materialTelephone interviews Easy to interviewto be shown to respondents(e.g. product/serviceproposition, literature) Can administer a longerinterview as interviewer/respondent rapport built Can be costly and timeconsuming to completegeographically unclusteredsample Typically lower cost than f2finterviews Cannot easily show stimulusmaterial, though can emailinfo/link for return call Interview typically shorterthan f2f interviewsWidely used method ofcommunication among b2baudiences Back toContentsOnline interviews Can use multimedia/showPostal interviewsstimulus Can be a quick and cheapermedium for high volumes, butnot always so Open end responses arenot probed, so answers oftenlimited Response rates often lowerthan other mediums withinterviewer contact Cannot be sure who hascompletedEssentially self completionso needs to be relevantand engaging to get goodresponse rateInterview length should bekept to a minimum Can be inserted with other Not used as frequentlymarketing material Open ended responsesun-probed, so answers oftenlimitedRouting and otherinstructions clear enough forrespondents to follow Response rates lowerthan other mediums withinterviewer contactEssentially self completionso need to be relevantand engaging to get goodresponse ratesInterview lengths should bekept to a minimum

Research methods2.6 Types of quantitative study availableIndividually commissioned custom studies One client Data confidential to clientSyndicated studies Group of clients share findings – outputs are also not confidential Copyright remains property of agency Typically market studies – where competitor benchmarkingrequired Often involve large sample sizes or hard-to-reach respondentsor limited universes Costs are shared among subscribers so are often cost effectiveOmnibus studies Agency will interview a set number each week/month Questionnaires will have a set classification section Clients ‘buy questions’ in the questionnaire Omnibus studies can be a cheap and quick way of gettinganswers to a few quick questions. They are often used to generatedata for PR purposesBack toContents

Research methods2.7 Sampling considerationsBack toContentsNot every contact will result in an interviewPutting customers firstCan't do ite.g. away duringfieldworkDoesn't want toe.g. company policy Response rate

3 Questionnaire and interviewing3.1 Questionnaire design and ideal interview lengthA good questionnaire should engagethe respondent from the startthe scene; tell them why research is being Setdone and why their help is of valuehonest/transparent about the potential length Beof the interviewrepetitive questioning AvoidKeep a logical order from a respondent point of viewAvoid excessive classification (dowe have some of this informationalready, e.g. on database orsample?)Ensure respondents have achance to ‘have their say’Back toContentsInterviews vary dependent on:type Respondent Context/relevanceof relationship Closeness Medium (f2f can be longer)The shorter the better. Max lengths:30 mins f2f:20 mins Phone: Online: 20 mins

Questionnaire and interviewing3.2 Questionnaire design – asking the right questionsAre we asking questions that can be answered:Accurately – known factse.g. What was your age last birthday?Through memorye.g. How many times in the last month have you done X?Through a best choice of options, none of which maycorrespond precisely to the respondent’s view/behavioure.g. Which of these three statements comes closest todescribing your view ?Through estimation, guesswork or even speculationBack toContents

Questionnaire and interviewing3.2.1 Questionnaire design – asking the right questionsBack toContentsGolden rules Is it possible for the respondent to answer (in business-to-business research for example, can a single informantanswer all the different questions? Can a customer answer all questions if a financial advisor was used?)?Keep question concepts understandable with as little ambiguity as possible. A question should beinterpretable in only one wayQuestions should be clear and phrased in language appropriate to the respondent’s way of thinking/talking.Consider the appropriateness of question wording to the audienceOnly one question should be asked at a time – questions containing multiple concepts (e.g. How would yourate the accuracy and speed ?) rarely give sensible dataEnsure that respondents are not led to a particular answerTake care with sensitive subject areas, often placing them towards the endAppropriate answer options should be available that reflect the reality of the range of responses. Give theopportunity to decline to answer a questionWhere appropriate include standard questions or questions used on previous research – it gives comparabilityacross studies and can enhance the value of the data to the clientPilot if possible

Questionnaire and interviewing3.3.1 Scales, nets, top box and mean scoresScales Back toContents Scales should be typically balanced. However, some argue that an unbalanced scaleis suitable because:good should be the midpoint as most providers would want to be better than goodcustomer experience should usually be positive so emphasis should be on movingfrom goodSome researchers prefer 5 or 7 point verbal scales, or some prefer non-verbal scoresout of 10. There is much debate about which is best and there is no real black and whiteanswer. The best starting point is ‘can the respondent answer the question?’It is generally good practice to keep scales consistent throughout a survey sorespondents don't get confused 100%100%2480%60%3440%Excellent (100)Very good (75)Good (50)Fair (25)Poor (0)2080%860%3020%11Q1 2014* In reality, people are rarely neithersatisfied nor dissatisfied2380%2160%1120%110%7Q1 201410102320%7Extremely satisfied (100)Very satisfied (67)Fairly satisfied (330Neither nor (0)*Fairly dissatisfied (-33)Very dissatisfied (-67)0% Extremely dissatisfied (-100)40%40%230%100%0%75643Q1 201410987654321

Questionnaire and interviewing3.3.2 Scales, nets, top box and mean scoresNet Reason for Complaint, SME YE Q1 2014 NETSErrors/mistakesCustomer ServiceChargesOnline bankingBranch RelatedPoor service (operational)LendingInternational servicePoor communicationRelationship ManagerMoney transmissionLack interest in my businessMerchant services/own cardsOtherBack toContents In pre-coded questions, you should have as many options as possible. Nets canbe created at the tables stage if needed (linked to client needs/groupings)Open-ended questions are usually recorded verbatim and can be coded at agranular level (to provide for more detailed understanding) and also at a net level(to allow for easier measurement)Total 2412988776654326Errors re account transactionsErrors re day to day adminOne or two major errors/mistakesAdmin errors/they lose thingsErrors/problems with BACS paymentsCharges raised in error/clarity of chargesMoney taken from account on error/withoutauthorisation/Paid into wrong account/Lostmoney/cheques paid into accountSilly/minor errors/mistakes4%1% 0.5%4%2%7%6%1%They had a bigbanking error aroundsystem failure Fees charged in errorTransaction in theaccount, deductionmade in error

Questionnaire and interviewing3.3.3 Scales, nets, top box and mean scoresTop box 100%80%60%box scores are often used by clients to target improvement because: TopGoalsare often for best in class (i.e. better than good) Staff betterunderstand the lexicon ‘Excellent/very good’ rather than a mean score 129333539Good (50)4850Fair (25)5832283120%Excellent (100)Very good (75)2740%0%Back toContents262319127165Q1 2013Q2 2013Q3 2013156113Q4 2013Q1 2014Poor (0)

Questionnaire and interviewing3.3.4 Scales, nets, top box and mean scoresMean scores 100%Back toContents Mean scores are a calculated average score. They can be calculated on a scale usedin the questionnaire or sometimes you can apply a value to the respondent's answer(e.g. excellent 100, very good 75, etc.)Mean scores take into account the full range of answers across the scale and providea single figure, which can be useful for tracking changeThe danger with using mean scores is that they can hide polarised views 64Mean scores64Excellent (100)2180%Very good (75)Good (50)51Fair (25)Poor (0)60%40%3716420%0%32027139AB

4 Tables, weighting, sig testing4.1 How to check tables (agency side)Back toContentsAny tables you receive from an agency should be fully checked. However, there may be occasionsyou have conducted in-house research. Therefore, when checking tables, for all tables you should: Check that each table you have specified is in fact presentCheck that its title, base and side headings are correctly described and that there are no spelling errorsCheck against the hole count to ensure the base is correctly definedCheck against the hole count to ensure that the side headings have been correctly defined. Ensure that theyadd up to 100% for single code questions and to at least 100% for multi-codesCheck that the correct breaks have been applied to the table and that they are correctly labelledRefer to the hole count to check that the breaks have been correctly defined. Make sure that whereappropriate they add across to 100%Check breaks against their corresponding table (to ensure for example that if frequency of use is used as abreak it corresponds with the frequency of use table)LOOK AT THE TABLE AND MAKE SURE IT MAKES SENSE Where possible, check that the table corresponds with previous or subsequent tables (e.g. spontaneousawareness and prompted awareness equals total awareness)You should also check any data that is derived from the sample – ask data processing if you don’t know whereto find the sample information

Tables, weighting, sig testing4.2 How to read tablesCharterhouse UK Business Confidence Survey 2014 (Survey Period: March - May 2014) - BC questionsTable 1Back toContentsQuestion/titleBaseBC1. How do you feel about the current health of the UK economy?Base: All answeringCross breakMain BankTotalBarclays BoSClydesdale Co-opHSBCLloyds NatWest RBSSub breakSignificance Level: 95%abcdefghUnweighted base24185051093645396380447220Total baseWeighted base3492112756284 1134573178282394607424 611611 594570 198932Extremely positive (100.0) 169038391124162126516663714928129 336978732Filtered base5%5%4%4%2%6%5%6%4%Fairly positive (75.0)2035703 438141781291147639969 375616 364800 334145 106293Column %58%58%69%36%49%62%60%56%53%cmacdghjmchjmcmcmSide labelNeither positive811006182378 196471480734556 111678 147131 137720 47546nor negative (50.0)23%24%17%47%42%18%24%23%24%Mean score pointeabefghijklabefghijklNet: excellent/very goodFairly negative (25.0)3728437650796054234467061996 58627 765852825411%10%8%13%6%10%10%13%14%Mean scoreExtremely negative (0.0)1035232014519140153420985 12924 1242381083%3%2%0%2%3%2%2%4%Standard error (used in significancetesting)NetPositive2204741 477253 822911274041635412765 392929 367842 115025Significance testing63%63%73%40%51%68%64%62%58%Significance testing at the 95% level is applied tocmcdghjlm cdhjmcmcmcthe tables. Each column is given a letter and thenNegative47636596652115194234620482981 715518900836362compared against each other. The letters shownunderneath denote where there is a statistically14%13%10%13%8%14%12%15%18%significant difference to other columnsfiMean score62.863.266.157.760.864.263.862.660Small base warninglmchlmhlmhlmlmIf a base is under 50 it will be denoted by a *Standard error0.440.941.813.122.661.111.041.021.56"Columns Tested: a,b,c,d,e,f,g,h,i,j,k,l,m - a,b,c,d,e,f,g,h,i,j,k,l,m,n,o - a,b,c,d,e,f - a,b,c,d,e,f,g,h,i - a,b - a,b,c,d"* small baseSurvey Period: March - May 2014 Charterhouse Research

Tables, weighting, sig testing4.3 Base sizesSmall base sizesQuotasBack toContents We should always be cautious of trying to segment the sampletoo much as results will become unreliable As a rule of thumb, bases under 50 should be treated with caution;bases under 20 should be considered more as qualitative indicativefindings With specialist audiences (e.g. businesses) we are often restrictedto using relatively small sample sizes (due to small universes andbudget restrictions). Where this is the case, we look for supportivedata within a data set and other client information, rather thanrelying purely on statistical observationsTo enable a large enough base for analysis at a sub grouplevel we often set quotas. This can result in certaingroups being over or under-represented at a total level.Thus, weighting is sometimes applied to make sure theresults are truly representative of the whole market

Tables, weighting, sig testing4.4 WeightingBack toContentsTotal respondents – 1,000Quotas setIn reality (% oftotal universe)WeightingfactorSector A –200 interviewsSector B –200 interviewsSector C –200 interviewsSector D –200 interviewsSector E –200 interviews20% of total20% of total20% of total20% of total20% of totalSector A –32%Sector B –25%Sector C –16%Sector D –8%Sector E –19%Sector A – 1.6Weighted base 320Sector B – 1.25Weighted base 250Sector C – 0.8Weighted base 160Sector D – 0.4Weighted base 80Sector E – 0.95Weighted base 190

Tables, weighting, sig testing4.5 Significance testingBack toContentsConfidence intervals The confidence interval is the range around the survey percentage or mean score into which the true figure forthe market as a whole is likely to fallThe 95% confidence interval means there is a 95% chance the true figure is within this range – and only a 5%chance it is outside this rangeFor percentages, the range is determined by the percentage and the sample base – the bigger the sample, thesmaller the rangeFor mean scores, the range is also affected by how clustered the scores are – the more clustered, the smallerthe standard error, and the smaller the rangeThe minimum significant difference between two percentages or mean scores is the difference required to be95% sure there is a real difference in the market as a wholeInteractive charts –enter figures intowhite cells

StatsCheckerBack toContentsDownload Charterhouse's significance testing appBack toContentsCharterhouse Research has developed StatsChecker, a FREE app for iOSand Android that enables clients to quickly and simply check whether surveypercentage, mean score or net promoter score (NPS ) differences arestatistically significant Populate the topfour cells – the %scores and thesample sizes tobe checkedThe app will calculatesignificance at a 95%confidence levelTest percentages, meanscores and net promoterscores (NPS )

5 TimelinesBack toContentsAn agency would turn around a typical ad hoc project in about six weeks, but taking into accountinternal project briefing and setup, it is wise to allow about nine 9Internal briefing meetingAn idea of approach and scopeIssue brief to agencyAgency proposalNear confirmation of approach and scopeAgency briefing meetingConfirmation of approach and scopeSet up, sampling, recruitmentFieldworkAnalysis & presentation(note: all projects are different and each will afford their own specially designedtimetable)

6 OutputsBack toContents Have you got your story straight?Ask yourself the “so what?” questionAny recommendations or questions to put forward?Do the findings have wider relevance? If so, for whom?Test hypothesis with key stakeholders before sharingwidelyA picture is worth a thousand words.but headlines are just as importantContext – will the content be delivered face-to-face, orread standalone?Consider the voiceover/amount of detail required,which may differ accordinglyAttention to detail/consistencyCareless mistakes sow the seeds of doubt!Labelling/keys/scales/appropriate chart types forthe data presentedBrand guidelinesFooters – source/question wording/base size

Charterhouse Research Ltd, 68 Lombard Street,London, EC3V 9LJ020 7868 1710www.charterhouse-research.co.uk

Charterhouse Research is pleased to bring you its Fundamentals of market research techniques guide. The book is intended to be a basic step-by-step guide to market research techniques, designed for new-to-research client-side research

Related Documents:

Pass Google ADWORDS-FUNDAMENTALS Exam with 100% Guarantee Free Download Real Questions & Answers PDF and VCE file from: . A key benefit of My Client Center (MCC) is that it allows: . Latest Google exams,latest ADWORDS-FUNDAMENTALS dumps,ADWORDS-FUNDAMENTALS pdf,ADWORDS-FUNDAMENTALS vce,ADWORDS-FUNDAMENTALS dumps,ADWORDS-FUNDAMENTALS exam .

26 FUNDAMENTALS OF MILLING – CONVERSATIONAL PROGRAMMING HEIDENHAIN 2 CNC fundamentals 2NC fundamentals C 2.1atums D Workpiece datum, machine datum, reference point 2. CNC fundamentals Datums Explain t

El Salvador Market entry: 2005 Units: 89 Costa Rica Market entry: 2005 Units: 217 Nicaragua Market entry: 2005 Units: 86 Honduras Market entry: 2005 Units: 81 India Market entry: 2009 Units: 20 Africa Market entry: 2011 Chile Units: 396 Market entry: 2009 Units: 404 Brazil Market entry: 1995 Units: 557 Argentina Market entry: 1995 Units: 105 As .

Market research pertains to research in a given market. As an example, a firm may conduct research in a target market, after selecting a suitable market segment. In contrast, marketing research relates to all research conducted within marketing. Thus, market research is a

Measurement and Scaling Techniques Measurement In Research In our daily life we are said to measure when we use some yardstick to determine weight, height, or some other feature of a physical object. We also measure when we judge how well we like a song, a File Size: 216KBPage Count: 23Explore further(PDF) Measurement and Scaling Techniques in Research .www.researchgate.netMeasurement & Scaling Techniques PDF Level Of .www.scribd.comMeasurement and Scaling Techniqueswww.slideshare.netMeasurement and scaling techniques - SlideSharewww.slideshare.netMeasurement & scaling ,Research methodologywww.slideshare.netRecommended to you b

Image restoration techniques are used to make the corrupted image as similar as that of the original image. Figure.3 shows classification of restoration techniques. Basically, restoration techniques are classified into blind restoration techniques and non-blind restoration techniques [15]. Non-blind restoration techniques

research objectives and the research questions. How to define a problem that can be solved by market research. How to write a brief for a market research project. What to expect in a market research proposal and how to choose between alternative proposals. Decisions that can be guided by market research

2 Korean Language Korean is an agglutinative language in which “words typically contain a linear sequence of MORPHS ” (Crystal, 2008). Words in Korean (eojeols), there-fore, can be formed by joining content and func-tional morphemes to indicate such meaning. These eojeols can be interpreted as the basic segmenta-tion unit and they are separated by a blank space in the Korean sentence. Let .