Market Research Glossary

A/B testing compares two versions of something, such as a web page, an email or a price, by showing each to a randomly assigned group and measuring which performs better against a defined outcome.

Because allocation is random, any difference in outcome can be attributed to the change itself rather than to differences between the two audiences.

It sits within quantitative research, since the result is a measured difference rather than an explanation of it.

Accompanied shopping involves accompanying consumers as they interact with a brand, before, during or after purchase. Another term for it is the accompanied shopping trip, or AST.

The researcher may accompany the shopper in store, online, during a phone conversation or in a face-to-face meeting. The shopper may be observed passively or questioned actively as the shopping activity unfolds.

The method gives a fuller understanding of the consumer’s conscious and subconscious needs, expectations and behaviours at each step of their journey. It also allows the experience the consumer wants to be compared with the one the brand is actually providing.

Brandspeak runs accompanied shopping as part of its qualitative research programmes.

An accompanied surf is the observation of consumers as they interact with a brand online, in their own way rather than against a set task. Observation may be passive or may involve questioning as the session takes place.

Accompanied surfing is related to, but distinct from, usability testing. A usability test sets defined tasks and measures whether the interface supports them; an accompanied surf follows the route the user would have taken anyway.

Accompanied surfs may take place in consumers’ homes or in configured viewing facilities. Eye-tracking is often added to show where attention falls and to identify issues the consumer may find hard to put into words.

The method enables B2C and B2B brands to make adjustments that improve the user experience, and with it rates of satisfaction and uptake.

Brandspeak uses it within customer experience research.

Acquiescence bias is the tendency for respondents to agree with a statement regardless of its content, often to be agreeable or to move through a survey quickly.

It inflates agreement scores. It is reduced by balancing positively and negatively worded statements, and by avoiding phrasing that signals the answer expected.

Controlling for it is part of questionnaire design in quantitative research.

Ad concept testing evaluates initial advertising concepts, or more finished executions, using either qualitative or quantitative approaches. It assesses relevance, understanding, impact, appeal and potential longevity.

It differs from concept testing proper, which evaluates a product or service idea rather than the advertising for it. Brandspeak covers both within its advertising development research.

Ad hoc research is a one-off study designed to answer a specific set of questions at a particular point in time. By contrast, ongoing or longitudinal programmes are designed to assess change over time.

By its nature, most qualitative market research is ad hoc.

Most of Brandspeak’s market research is commissioned this way.

Ad recall research measures the extent to which a respondent can correctly recall the content of a specific advertisement or set of advertisements.

It can be conducted on an aided or unaided basis. Aided recall prompts respondents with the advertisement when the focus is on recall, understanding and relevance. Unaided recall measures knowledge of the brand and the advertisement without prompting, which is the harder test because it relies on top-of-mind recall.

It forms part of advertising development research.

Advertising claim substantiation is the testing of product performance in order to demonstrate that a claim made in advertising can be supported.

In the UK, claims must be capable of substantiation before they are published, which means the research usually has to be designed and documented to a standard that will stand up to challenge.

Brandspeak covers claim testing within advertising development research.

Advertising effectiveness research determines the extent to which an advertisement or campaign has met its marketing objectives.

It may be conducted after a campaign has run, or continuously through brand tracking, which measures the cumulative effect of advertising on brand measures over time.

This is the initial testing of different brand positioning statements with the target audience, in order to understand which will resonate most strongly in advertising.

See advertising development research.

Advertising research is a broad term covering qualitative or quantitative research on advertising that is either in development or already launched, whether on television, in cinema, or through radio, press, online or outdoor media.

It is undertaken to refine strategy or execution, and after launch to monitor whether campaign objectives are being met. Post-launch research can also inform further edits.

Brandspeak’s offer in this area is advertising development research.

Advertising tracking is the periodic measurement of the appeal, relevance and impact of individual advertisements over time.

It is narrower than brand tracking, which measures the health of the brand itself rather than the performance of particular executions. The two are often run together.

Brandspeak runs it alongside brand tracking.

Anchoring is the tendency to rely too heavily on the first piece of information encountered when making a judgement.

In pricing research it matters a great deal. The first price a respondent sees shapes what they go on to consider reasonable, which is why price questions are asked in a controlled order and why anchors are rotated between respondents.

It is one reason pricing questions are sequenced carefully in quantitative research.

An animatic captures and communicates an advertising idea during its earliest stages of development. Animatics are usually sketches or cartoons, mounted on board or combined into short videos supported by voiceover or soundtrack.

They are often used in advertising development research to explore the underlying creative idea. Their unfinished nature requires careful handling by the research agency when they are shown to respondents, since reactions to the execution can be mistaken for reactions to the idea.

An alternative is the stealomatic, which uses existing film clips to convey the intended style of the proposed advertisement.

Animatics are used in advertising development research.

Attribute analysis is a technique used to develop lists of the characteristics, uses or benefits associated with a product category, which can then be tested for importance and for how strongly each is associated with individual brands.

It is often a precursor to brand tracking, where the attributes become the tracked measures.

Attribution modelling assigns credit for a conversion across the marketing touchpoints that preceded it.

It works well for measurable digital activity and poorly for brand-building activity whose effect is delayed and diffuse. That limitation is one reason brand tracking and marketing mix modelling are used alongside it rather than replaced by it.

Brandspeak measures the brand effect that attribution misses, through GrowthTrack.

Audience research measures how people respond to programming or advertising: who is reached, what is remembered, and how the content is received.

It may be conducted in a viewing environment, through a panel, or through recall and recognition questions asked after exposure.

See advertising development research.

Availability bias is the tendency to judge an event as more likely because examples of it come to mind easily. It was identified by Amos Tversky and Daniel Kahneman in 1973.

When a serious road accident receives heavy media coverage, those exposed to the coverage are more likely to imagine such an event happening to them, which is one reason insurance is often bought in the period after a widely reported incident.

It is one of the cognitive biases considered in brand strategy consulting.

B2B research is research undertaken with business customers rather than consumers. The term stands for business to business, in contrast to B2C, or business to consumer.

It would apply, for example, to a vehicle rental firm researching its corporate customers, or a bank seeking the views of its small business customers.

B2B samples are smaller, harder to reach and often involve a buying committee rather than a single decision maker, so sampling is tighter and base sizes are lower than in consumer work. Brandspeak has an established B2B research practice.

B2C research is any form of research in which a business or organisation seeks the views of consumers on an aspect of its brand or its delivery. The term stands for business to consumer.

Brandspeak conducts B2C work across FMCG, retail and other consumer sectors.

Behavioural economics explores what causes people to make decisions that are economically irrational, and so undermine the predictions of established economic models. It is sometimes shortened to BE.

Heuristics are a central example: the mental shortcuts that people use to simplify decision-making.

Its findings inform how Brandspeak approaches brand strategy consulting.

Behavioural segmentation groups customers by what they do rather than by who they are: what they buy, how often, through which channel, and how they respond to price or promotion.

It is usually the most commercially actionable basis for segmentation, because the behaviour that defines each group is already observable in the data.

Bias is systematic error that pushes results consistently in one direction. It is introduced by how a sample was drawn, how a question was worded, how an interview was conducted, or by who chose to respond and who did not.

Bias is distinct from sampling error, which is random and can be quantified. Bias cannot be reduced by increasing the sample size, because a larger sample simply measures the same distortion more precisely.

In qualitative research the question is more complex, because the researcher is part of the instrument. Bias can arise from the moderator’s own views, from how questions are framed in the moment, or from an unbalanced group.

Managing it is central to both qualitative and quantitative research design.

Big data refers to very large, complex data sets, both structured and unstructured, drawn together from a wide variety of sources within an organisation, such as market research data combined with operational data.

The insight that can be drawn from it is considerable, but so are the challenges of capture, storage, analysis, transfer and visualisation.

It is often combined with survey data in customer segmentation research.

Biometrics is the measurement and statistical analysis of physical and behavioural characteristics. In market research, certain biometric measures are used to capture physiological responses to stimulus such as advertising, shelf layouts, product concepts and packaging.

The measures most often referred to are eye-tracking, which records where the gaze falls and for how long; facial coding, which classifies facial expressions into a smaller set of inferred emotions; and electroencephalography, or EEG, which measures electrical activity in the brain in response to stimulus.

Claims made for these techniques vary widely in how well they are evidenced, and they are generally more useful as a supplement to questioning than as a replacement for it.

Related techniques are described under neuromarketing research.

A brand is a name, term, design, symbol or set of words configured to create an image that identifies one seller’s product or service as distinct from the competition.

Over time that image also becomes associated with the product’s features, benefits and other qualities, cementing its position in the consumer’s mind.

Brandspeak’s brand work runs from brand strategy consulting through to measurement.

Brand architecture is the structure of the brands an organisation owns and the relationships between them: which carry the parent name, which stand alone, and how each is signposted to the customer.

Research tests whether the intended structure is understood, and whether it helps or hinders choice. A structure that makes sense on an internal chart often makes none to a buyer in a shop.

Brand associations are the rational and emotional associations consumers hold in relation to a brand, derived from sources including advertising, word of mouth, packaging, pricing and actual use of the product.

Research identifies these associations and establishes their relative importance. Once the rational and emotional, prompted and spontaneous associations are known, they inform brand repositioning, advertising development and new product development, as well as communications and customer experience strategy.

They are measured over time through brand tracking.

Brand awareness is the extent to which a brand comes to mind, or is recognised, when a category or a need is considered.

It is measured two ways. Spontaneous awareness asks which brands come to mind unprompted. Prompted awareness asks which are recognised from a list. Spontaneous awareness is the harder and the more meaningful measure, because it reflects whether a brand is present at the moment of choice.

It is a core measure in brand tracking.

Brand equity is used in two distinct senses, and it is worth separating them.

Consumer-based brand equity is the strength of what people know, feel and expect of a brand, and the extent to which that makes them more likely to choose it or to pay more for it. This is what research measures.

Brand valuation is the financial worth of the brand as an asset on the balance sheet. This is an accounting exercise rather than a research one, though it often draws on research inputs.

Consumer-based equity is measured through brand tracking.

Brand essence is a distillation of a brand’s intrinsic characteristics into a short statement that encapsulates what it stands for. A clearly defined brand essence acts as a template in the development and execution of marketing strategy.

It is developed through brand development research.

A brand extension is a new product marketed under an existing brand name, drawing on the credibility the brand already holds in order to enter a new category or segment.

Extensions are tested through innovation and NPD research.

Brand health is the overall condition of a brand in its market, read across awareness, consideration, usage, perception and loyalty rather than from any single measure.

It is tracked over time so that a decline becomes visible while there is still time to respond to it. Brand tracking is the method by which it is measured.

Brand image is how a brand is perceived by consumers, which is not always the same as how it is portrayed by its owners. Research exists in large part to establish the size of that gap.

It is measured through brand tracking.

Brand loyalty is the extent of consumer preference for one brand relative to others, expressed through repeat purchase, resistance to switching and willingness to recommend.

Loyalty is one of the measures carried in GrowthTrack.

Brand personality is what people think and feel about a brand, consciously and subconsciously, described in terms of human personality traits. It can be developed or altered through marketing activity and is used to build engagement and attachment.

A brand’s personality should support its proposition and its positioning.

It is explored through brand development research.

A brand portfolio is the full set of brands an organisation owns. Portfolio research examines whether the brands within it compete with one another, whether each earns its place, and where gaps exist that a new or acquired brand could fill.

Portfolio questions are addressed through brand strategy consulting.

Brand positioning is the distinctive place a brand occupies in the target customer’s mind, relative to its competitors. A positioning should give the customer a compelling reason to choose the brand over the alternatives.

In Positioning: The Battle for Your Mind, Al Ries and Jack Trout offer a simple template: [brand] is [kind of product or service] for [kind of people]. The pasta manufacturer De Cecco used this to arrive at a positioning as the premium pasta for serious chefs.

In ZAG, Marty Neumeier proposes a longer form covering category, point of difference, customer, geography, need state and the underlying trend. Applied to Harley-Davidson, it produced a positioning built on big, loud motorcycles for riders seeking freedom and belonging, later distilled into the line American by birth, rebel by choice.

Brandspeak develops and tests positioning as part of its brand development research.

Brand price trade-off is a pricing research approach that identifies the price consumers are prepared to pay for a specific brand, and the effect different price points have on brand choice. Because it assesses price elasticity at brand level it also functions as a measure of brand equity.

Respondents consider a basket of brands at a series of price points, selecting their preferred brand at each. Repeating the exercise across price points reveals the point at which preference shifts.

Brand research is a catch-all term for any research carried out in relation to a brand.

It may be early qualitative work identifying the strategic building blocks of a new brand, later qualitative or quantitative work developing elements of the marketing mix such as name, packaging and pricing, or post-launch tracking that monitors performance against objectives.

See brand development research and brand tracking.

Brand salience is how readily a brand comes to mind in a buying situation.

It is a measure of noticing rather than liking. A brand can be well regarded and still rarely thought of at the point of choice, which is why salience is tracked separately from preference and why it is more closely linked to market share.

Salience is tracked within GrowthTrack.

Brand tracking is continuous or wave-based research carried out for both B2C and B2B brands, designed to monitor a brand’s performance and health over time and to show the effect of advertising and other marketing activity.

Key metrics typically cover awareness, consideration, usage and loyalty, as well as attitudes and perceptions. The outputs inform both short-term tactics and long-term strategy.

B2C trackers are usually conducted online. For B2B brands, tracking is more often undertaken by telephone or face to face, because the audiences are smaller and harder to reach.

Brandspeak has run brand trackers for over 20 years and in more than 30 markets. GrowthTrack is its brand tracking programme, which links brand measures to acquisition, retention and switching.

CAPI is a survey administered in person by an interviewer using a computer-based questionnaire, with the respondent’s answers typed directly into the system.

The software checks for invalid responses, routes the respondent through the questionnaire according to their answers, and will not accept responses outside the limits set.

It is one of the data collection modes used in quantitative research.

Card sorting, sometimes called mapping, is a qualitative technique in which respondents group a set of elements according to criteria of their own choosing or criteria supplied by the moderator.

Potential product features and benefits may be written on individual cards, with respondents asked to sort them into groups that make sense to them, or into categories such as more and less important.

It is a useful way of helping respondents structure and articulate thoughts they would struggle to express in open discussion. It should not be confused with customer journey mapping, which is a different exercise entirely.

It is a qualitative research technique.

CASI is an interview method in which the respondent completes a computer-based questionnaire themselves rather than answering an interviewer. It is often used where subject matter is sensitive and an interviewer’s presence would encourage socially desirable answers.

See quantitative research.

Category entry points are the situations, needs and occasions that prompt someone to think of a category: the moments at which a purchase begins.

Research identifies which of them a brand is linked to. Growth usually comes from being associated with more entry points rather than from being preferred within one, which is the practical implication of work on mental availability.

They are measured within GrowthTrack.

CATI is interviewer-administered quantitative telephone surveying using a computer-based questionnaire. It can be run from a call centre or by interviewers working from home, for both B2B and B2C research.

The software presents the questions on screen and records answers directly, routing the respondent automatically according to their answers and removing a common source of human error.

Brandspeak uses CATI particularly for B2B research.

The chi-square test establishes whether there is a statistically significant relationship between two categorical variables, such as whether brand preference differs by region.

It is one of the significance tests applied in quantitative research.

Cluster analysis is a class of techniques used to group individuals within a population according to how similar they are across multiple characteristics. It is also referred to as classification analysis or numerical taxonomy.

It is the usual means of building a segmentation from survey data.

Co-creation is the rapid identification, capture and creative exploration of ideas or concepts, typically within a focus group, workshop or online community. Sessions may focus on new brands, products, services or customer experiences.

Sessions often include client representatives alongside researchers and consumers, so that ideas can be guided towards a useful outcome. Workshop co-creation is highly collaborative; online co-creation allows larger numbers to contribute and build on one another’s ideas.

A code of conduct sets out the ethical and professional standards within which researchers operate, detailing their obligations to clients and to participants.

In the UK the relevant code is that of the Market Research Society. Brandspeak is an MRS member and works to its Code of Conduct.

Cognition is the process by which knowledge and understanding are acquired through experience, the senses and thought. It may be conscious or unconscious.

Cognitive biases are decision-making shortcuts that enable fast, efficient judgements. They may also arise from emotional predisposition, social influence or the limits of attention and memory.

They are efficient but systematically error-prone. Amos Tversky and Daniel Kahneman conducted the foundational research in this area during the 1970s.

They inform how stimulus is designed and presented in qualitative research.

Competitor analysis may be ongoing or ad hoc. Ongoing analysis is usually built into a brand’s own tracking study, reviewing its performance against key competitors on awareness, consideration, usage, loyalty and perceptions.

Ad hoc competitor analysis is more often commissioned when a new market is being considered for entry, or when a competitor begins to make inroads into an established brand’s market.

It will typically address a competitor’s target market, proposition and positioning, product features and benefits, communications strategy by channel, offers and promotions, pricing, and the attitudes and perceptions attached to the brand.

Brandspeak offers this as competitor analysis.

Concept testing assesses consumer reactions to a new idea, typically for a product or service, while it is still expressed in words rather than built.

Qualitative concept testing suits the exploratory phase, when the question is why an idea works or fails. Quantitative concept testing suits the choice between several candidates, when the question is which to progress.

It establishes whether an idea has enough merit to pursue and what needs to change before development continues. In a quantitative test, each respondent may see one concept, which is a monadic design, or several in turn, which is sequential monadic.

Brandspeak runs concept testing within its communications research practice.

A confidence interval is the range within which the true value in the population is likely to sit, given a result from a sample.

A survey finding of 40% on a sample of 1,000 carries an interval of roughly three percentage points either side at the conventional 95% level. Reporting a figure without its interval implies a precision the sample does not support.

Intervals are reported alongside findings in quantitative research.

Confirmation bias is the tendency to seek out or interpret information in a way that confirms beliefs already held, and to discount information that challenges them. The term was coined by the British psychologist Peter Wason in the 1960s.

It is one of the strongest arguments for independent research, since a team testing its own idea will tend to find support for it.

It is among the reasons organisations commission independent research.

A conflict interview recruits respondents holding different or opposing attitudes on a specific issue, so that it can be explored from more than one perspective within a single session.

The approach surfaces the strength with which views are held and can identify common ground between otherwise opposed groups.

It is a qualitative research technique.

Conjoint analysis establishes how consumers make choices when several factors have to be considered jointly, which is where the name comes from. It quantifies the relative value attached to individual product attributes, including price.

It is used to determine what features a new product or service should have and how it should be priced, by observing the trade-offs people make rather than asking them directly what matters.

Brandspeak delivers conjoint within its pricing and innovation and NPD research.

A consumer insight is an understanding of why people behave as they do that is specific enough to act on.

A finding describes what happened. An insight explains it well enough to change a decision. Most research produces a great deal of the first and rather less of the second.

Turning findings into insight is the purpose of Brandspeak’s market research.

Convenience sampling recruits whoever is easiest to reach.

It is quick and inexpensive, and it cannot support claims about a wider population, because people who were easy to reach are rarely typical of it. It has a place in exploratory and pilot work, and none in measurement.

Sampling approaches are covered under quantitative research.

Copy development research, also called pre-testing, identifies the reaction of target consumers to early-stage advertising concepts or copy.

See advertising development research.

Copy testing establishes the level of understanding, impact, awareness and credibility that an advertisement generates.

Brandspeak covers it within communications research.

Correlation is a measure of the degree to which two variables move together. Variables are positively correlated when high values of one accompany high values of the other, and negatively correlated when high values of one accompany low values of the other.

A correlation says nothing about whether one variable causes the other. Two measures can correlate strongly because both respond to a third factor, or by coincidence. Correlation analysis and simple linear correlation describe the statistical techniques used to quantify the relationship.

Correlation and regression underpin the driver analysis used in GrowthTrack.

Creative development research is qualitative advertising research that examines the strength of one or more advertising concepts in order to identify candidates worth developing further.

See advertising development research.

Cross tabulation examines responses to one question against responses to one or more others, which is how most survey analysis is conducted in practice.

Cross tabulations are a standard deliverable in quantitative research.

Cross-cultural research is conducted in two or more countries, or across two or more cultural or ethnic groups, and must account for differences in language, norms and market maturity when results are compared.

Brandspeak conducts it as international market research.

Customer effort score measures how much work a customer had to do to get something done, usually through a single question asked after an interaction.

It predicts repeat custom better than satisfaction in service settings, because effort is remembered where satisfaction fades.

It is used within customer experience research.

Customer experience mapping is the process of visually capturing every interaction a customer has with a brand. Individual interactions are known as touchpoints.

For each touchpoint the nature of the experience is captured and described, with particular attention to the pain points causing dissatisfaction and, in some cases, outright rejection of the brand.

The map allows those pain points to be prioritised and addressed, so that satisfaction, loyalty and retention can be improved.

Brandspeak offers this within customer experience research.

A customer journey map tells the story of the end-to-end customer journey across all touchpoints, potentially from initial awareness through to first contact, purchase, after-sales support and renewal or repurchase.

It sets the experience the customer would like to receive against the experience the brand has actually laid out, and identifies where the two diverge. Those divergences can then be researched qualitatively or quantitatively to establish their nature and their effect.

The findings form the basis for prioritising remedial action.

See customer experience research.

Customer lifetime value is the total profit a customer is expected to generate over the whole of their relationship with a brand.

It is the figure that justifies spending to acquire or retain them. Research contributes the behavioural inputs: likely tenure, purchase frequency and switching risk.

Its behavioural inputs come from customer experience and segmentation work.

Customer retention is the proportion of customers who remain with a brand over a defined period.

Research explains the number rather than reporting it, identifying which experiences cause people to leave and which hold them, so that effort goes to causes rather than symptoms.

Retention drivers are identified through customer satisfaction research.

CSAT is the proportion of customers reporting that they are satisfied, usually from a single question asked after an interaction or at intervals through a relationship.

It is simple to collect and simple to compare over time, but it reports a level without explaining it, which is why it is normally paired with diagnostic questions. Customer satisfaction research more broadly uses those diagnostics to identify which aspects of the experience are driving the score.

Brandspeak runs customer satisfaction research using CSAT alongside diagnostic measures.

Customer segmentation divides customers into groups based on similarities relevant to marketing, such as age, life stage, interests or spending habits. It can also be built on attitude or behaviour, which is usually more useful.

In business segmentation the criteria are more likely to be sector, turnover and number of employees, sometimes described as firmographics.

Brandspeak builds and validates segmentations through its customer segmentation research.

CX research identifies and captures the customer’s experience of a brand across one or more touchpoints. CX stands for customer experience.

See customer experience research.

A CX survey quantifies the nature of the customer experience a brand is delivering, usually at defined points in the journey rather than as a single overall measure.

See customer experience research.

Data analysis is the process of collating, filtering and examining data in order to identify patterns and findings that answer the question the research was commissioned to address.

It is the stage at which quantitative research becomes usable.

Data collection is the process of gathering and collating research data for analysis, covering fieldwork, data capture and the checks applied before analysis begins.

See quantitative research and qualitative research.

Data processing covers the classifying, summarising, recording, calculating, sorting and storing of data so that it can be accessed and analysed. In survey research it also includes coding open-ended responses and applying any weighting.

See quantitative research.

Data visualisation is the presentation of findings in visual form so that a pattern can be seen rather than worked out.

In research it is the difference between a table a reader has to interpret and a chart that makes the point on sight, and a good deal of insight is won or lost at this stage.

It shapes how Brandspeak reports brand tracking results.

Deliberative research is often used in public consultation on policy. Participants are given a range of issues and scenarios to explore and debate before considered opinions are sought at the end, on the basis that an informed view differs from a first reaction.

It is a qualitative research method.

A depth interview is a qualitative interview conducted with a single respondent. It is used where subject matter is sensitive or complex, and so unsuited to a group, or where the respondent is hard to bring together with others.

Depth interviews are often conducted in the respondent’s own home, particularly where there is an observational element. Brandspeak runs them as part of its focus groups and depth interviews practice.

Desk research uses publicly available information, purchased reports or data already compiled by the client, including operational data. It may be conducted as a precursor to primary research or to supplement it.

It is also referred to as secondary research.

A diary study asks participants to record behaviour as it happens, over days or weeks, rather than recalling it afterwards.

It suits anything habitual, frequent or easily forgotten, where recall is unreliable. It is now usually run through an app, with photographs and short videos alongside written entries.

Brandspeak runs diary work within ethnographic research.

Econometrics is the analysis of economic systems using statistical models, applying them to supply, demand and pricing data in order to estimate relationships and forecast outcomes.

It is related to the modelling used in GrowthTrack.

Elasticity measures the extent to which one variable shifts in response to movement in another. In pricing it describes how far demand changes when price changes, and it is the central output of most pricing research.

Elasticity is the central output of pricing work within quantitative research.

An employee value proposition sets out what an organisation offers its people in return for what it asks of them.

Research tests it much as a brand proposition is tested: whether it is believed by those inside, whether it is distinctive to those outside, and whether it influences the decision to join or to stay.

EVPs are tested using the same approach as brand development research.

ESOMAR is the global association for market, opinion and social research and data analytics, founded in 1948 as the European Society for Opinion and Market Research.

It sets international standards for research practice, publishes guidance used widely across the industry, and maintains a code jointly with the International Chamber of Commerce.

Ethnography is the observation of individuals or groups within the context of their own environment. A researcher may spend hours, days or weeks observing and interacting with participants in the course of their everyday lives.

It sheds light on needs, attitudes and behaviours that respondents would not report accurately in an interview, either because they have not noticed them or because they would describe them differently when asked.

It is often used to identify new product opportunities, to define criteria for product development, and to inform advertising and customer experience work.

Brandspeak offers auto-accompanied ethnography among its ethnographic methods.

An exit interview is conducted with a customer or employee who is leaving, in order to understand why. It is used by retailers and service businesses as well as in employment settings.

Exit interviews surface brand, product, experience or employment issues that were previously unknown, or give a different perspective on known ones, so that satisfaction and retention can be improved.

They are used within customer satisfaction research.

Eye-tracking is a behavioural measure of visual attention. Infrared sensors record precisely where a person looks, what they read, what they pass over, how long they spend on each element and the order in which their gaze moves.

It measures where attention falls rather than what happens in the brain, and it captures something respondents cannot report reliably, because people are poor witnesses to their own looking.

It suits both static and dynamic content: advertising, outdoor posters, packaging, websites and printed material. The technology is available in fixed and portable formats.

Brandspeak delivers eye-tracking through a specialist partner, most often within packaging research and communications testing.

A face-to-face interview is conducted in person rather than remotely by web, telephone or post. It remains the strongest method where stimulus has to be handled, where rapport matters, or where the respondent would not engage properly by other means.

See focus groups and depth interviews.

Firmographics are the characteristics used to describe and group organisations in B2B research: sector, size, turnover, growth stage, geography and ownership.

They are the business equivalent of demographics, and on their own they are usually a weaker predictor of buying behaviour than need, buying structure or where the organisation sits in its own growth cycle.

They are used in B2B research and segmentation.

A focus group is a moderated discussion convened for research purposes among a group of respondents, typically eight to ten, who are usually unknown to one another.

Groups take place in private homes, hotels or research facilities, or online. The discussion is led by a moderator working to a pre-prepared guide that is responsive to how the conversation develops.

They are used to explore and resolve questions about brands, products and services, advertising and other communications, and customer experience.

Brandspeak runs focus groups and depth interviews in the UK and internationally.

Gabor-Granger is a pricing technique in which respondents say whether they would buy at each of a series of prices, presented one at a time.

It produces a demand curve and a revenue-maximising price. It suits a single product with a known specification rather than a range of configurations, for which conjoint analysis is better suited.

It is one of the pricing techniques Brandspeak uses within quantitative research.

Gamification makes online surveys more engaging by making them more game-like, in response to falling completion rates and declining data quality in long or repetitive questionnaires.

Examples include sliders, drag-and-drop tasks and questions answered against the clock. The trade-off is that a more engaging format can also change how people answer, so it needs to be applied consistently if results are to be compared over time.

It is a design choice within quantitative research.

Generation X refers to those born between roughly 1965 and 1980.

Their formative years were shaped by recession and high unemployment in the 1980s, rising divorce rates and the growth of dual-income households, the end of the Cold War, and the arrival of home computing and cable television.

Commentators generally characterise the cohort as self-reliant and sceptical of institutions, including of advertising, rather than as idealistic. Generation X was followed by Generation Y, also called the millennials.

Generation Y, also known as the millennials, is the generation that followed Generation X. The dates are generally given as the early 1980s to the mid-1990s.

The cohort is more socially liberal than its predecessor and markedly more comfortable with technology, having come of age alongside the consumer internet. Its early working years coincided with the financial crisis of 2007 to 2009, which delayed home ownership and family formation relative to earlier cohorts.

Generation Z refers to those born between roughly 1997 and 2012, the first cohort with no adult memory of life before the smartphone.

Research with this group has to account for low tolerance of long surveys, high fluency in video and short-form content, and a stated concern with brand ethics that does not always translate into purchase behaviour. Generation Alpha, born from around 2013, follows.

Brandspeak researches younger audiences across FMCG and retail.

A hall test is a quantitative method used to evaluate brands, products, packaging and advertising in a controlled environment. It is particularly suited to physical stimulus, such as a new food or drink product.

Hall tests can cover 40 to 60 respondents a day depending on the complexity of the material and the number of interviewers. They are held where target consumers pass in numbers, with recruiters screening people before inviting them in.

Eye-tracking can be added where visual attention is the question. Brandspeak runs hall tests through established partners.

Hall tests are used in packaging research and product testing.

Heuristics are the mental shortcuts people use to make everyday decisions, enabling them to decide far more quickly than if they processed all the relevant information.

A simple example is the rule that expensive means good quality. Heuristics are efficient in daily life, but they also produce poor decisions when they cause relevant information to be ignored.

Thinking, Fast and Slow by Daniel Kahneman is the best known general treatment of the subject.

They inform how Brandspeak approaches brand strategy consulting.

A home use test places a product with participants to use in their own homes, in normal conditions, over a set period.

It gives a more realistic read than a central location test on anything whose performance emerges with repeated use, such as a cleaning product, a skincare range or a food item eaten more than once.

Brandspeak runs product placement work within innovation and NPD research.

In developing a new brand, product or service, marketers often commission exploratory research in order to form a hypothesis about how it should be configured. That initial work is usually qualitative.

The hypothesis is then tested and refined through a further round, usually quantitative, before further investment is committed.

It is the usual bridge between qualitative and quantitative research.

Implicit testing uses time-tracked response tasks to capture which associations come to mind quickly and which take deliberation.

Because responses are recorded against the clock, they are harder for a respondent to manage than answers on a rating scale. The method is most useful where stated and actual preference tend to diverge.

Brandspeak offers implicit testing using time-tracked tasks.

Incidence rate is the proportion of the general population that qualifies for a study.

It is the main driver of fieldwork cost, because a low incidence means screening many people to find each eligible respondent. This is why niche B2B audiences cost more per interview than general consumer samples.

It is a principal driver of fieldwork cost in quantitative research.

An independent variable is one that can be controlled or set. A dependent variable is one that responds to changes in it.

The elements of the marketing mix, price, product, place and promotion, are independent variables. Sales growth and average transaction size are dependent variables that respond to them.

See quantitative research.

Informed consent means a participant has agreed to take part knowing what the research involves, how their data will be used, and that they may withdraw.

It is a requirement of the MRS Code of Conduct and of data protection law. It is also a condition of the research being usable, since consent obtained improperly puts the resulting data at risk.

Brandspeak works to the MRS Code of Conduct across all market research.

In research, insight refers to the conclusions drawn from analysing qualitative or quantitative data. The term is often used loosely for any finding, but an insight properly explains behaviour well enough to change a decision.

See Brandspeak’s market research services.

International market research covers qualitative and quantitative, B2C and B2B research conducted across countries. Projects are often multi-market, with the same research repeated in several countries and the findings reconciled.

Conclusions must take account of market-specific social and cultural influences and of differing levels of market maturity. A question that works in one market can mean something different in another, which is why translation and back-translation matter.

Brandspeak conducts international market research in more than 30 markets through native-language partners.

ISO 20252 is the international quality standard for market, opinion and social research, covering how projects are specified, how data is collected and checked, and how records are kept.

Certification is held by the agency rather than by the individual researcher, and it is often asked for in public sector and large corporate procurement.

Standards and conduct are covered under Brandspeak’s market research services.

Jobs to be done frames demand around the progress a customer is trying to make rather than around who they are.

The question becomes what someone hired this product to do, which tends to surface competitors from outside the obvious category. A milkshake bought to occupy a commute competes with a banana, not with another milkshake.

The framing is used in innovation and NPD research.

Key drivers analysis is a modelling technique that examines the relationships between many factors in order to identify those with the greatest influence on an outcome.

It answers which activities are most likely to deliver a business objective, and which attributes matter most to customers, as distinct from which they say matter most.

Driver analysis is built into GrowthTrack.

Laddering is a qualitative technique requiring skilled moderation. It uses repeated questioning, in effect asking why several times over, to move beyond superficial rational responses and reach more deeply held motivations.

It produces a richer understanding of why a choice is made, which can then inform brand and advertising development.

It is a qualitative research technique requiring experienced moderation.

A Likert scale asks respondents to place themselves on a graded scale between two opposites, most often from strongly agree to strongly disagree. Five and seven point versions are standard.

Results are usually reported as a top-two-box figure, combining the two most positive responses, which is simpler to track over time than a mean score.

Scales of this kind are standard in quantitative research.

Longitudinal research monitors change in perception, behaviour or demand over time.

A study that follows the same group of people over a protracted period is called a cohort study, where participants share a characteristic such as year of birth.

Brand trackers are the most common form of longitudinal work. See brand tracking.

Margin of error is the amount by which a survey result is likely to differ from the true figure in the population, as a consequence of sampling rather than counting everyone.

It narrows as sample size grows, but by less than people expect: quadrupling the sample halves the margin. It is the reason small differences between waves of a tracker should not be over-read.

It should be read alongside any quantitative research finding.

Market entry research provides the information needed to make decisions about entering, expanding in or withdrawing from a market. It is sometimes combined with market sizing.

These studies typically blend qualitative and quantitative research with desk work using publicly available information and, where available, the client’s own data.

A market entry report may cover market background and size, growth trends, customer segmentation, the main players and their propositions, distribution and communications, and the gaps and opportunities that follow.

Brandspeak combines primary and secondary research for market entry work.

An MROC is a moderated online community in which a recruited group takes part in discussions and tasks over a period of weeks or months. It is also called an online research community.

It is undemanding of respondents, who dip in and out, and its duration allows a more complete exploration of opinion than a single session. It also lets the research team develop its thinking as the work progresses and adjust the discussion accordingly.

Participants respond to threads, upload video and other material, and interact with one another as well as with the moderator.

MROCs sit within qualitative research.

The Market Research Society is the UK professional body for research, insight and analytics, and the body whose Code of Conduct governs how research is conducted in this country.

The Code sets requirements on informed consent, the treatment of participant data, honesty in reporting, and the duty not to mislead participants about the purpose of research. Membership is held by individuals and by companies, and members agree to be bound by it.

Brandspeak is a member of the MRS and of the AQR, the Association for Qualitative Research.

Market segmentation divides a market into groups that differ in ways that matter commercially, so that each can be approached differently.

A segmentation is only useful if its segments can be identified in the real world, reached through available channels, and are large enough to be worth the effort. A segmentation that is elegant but cannot be found in a customer database will not be used.

Brandspeak builds and validates segmentations through customer segmentation research.

Marketing analytics is the practice of measuring and analysing marketing performance in order to improve its effectiveness and optimise return on investment.

See GrowthTrack for brand measurement linked to commercial outcomes.

Marketing mix modelling uses statistical analysis of historical sales and marketing data to estimate what each activity contributed.

It captures the effect of brand activity that attribution models miss, and it requires several years of reliable data, which is its main practical constraint. It is sometimes called media mix modelling.

Brandspeak measures the brand contribution through GrowthTrack.

MaxDiff asks respondents to choose the most and least important items from small sets shown repeatedly, producing a ranked list with clear distances between the items.

It avoids the central problem of rating scales, where respondents call almost everything important, and it is the usual choice for prioritising messages, features or benefits.

It is used to prioritise messages within communications research.

Mental availability is the likelihood that a brand is noticed or thought of in a buying situation.

It is built by linking a brand to many category entry points rather than by persuading people that it is better. The concept is most closely associated with the Ehrenberg-Bass Institute.

It is measured within GrowthTrack.

A mini-group is a smaller group discussion, usually involving four respondents. It is also called a quad, and it allows more depth per participant than a full group while retaining some group dynamic.

See focus groups and depth interviews.

Mobile market research is research in which the respondent takes part through a handheld device. It suits in-the-moment capture, such as photographing a shelf or recording a reaction immediately after an experience.

It is often used within ethnographic research.

In a monadic design each respondent evaluates a single item in isolation, rather than comparing it with alternatives.

It gives the cleanest read on each item, because the respondent is not making relative judgements, but it requires a larger total sample since each item needs its own group. Where several items must be compared economically, a sequential monadic design is used instead.

Brandspeak has written on monadic versus sequential monadic testing in more detail.

Multi-modal research asks respondents to describe brands in terms of the five senses: smell, taste, sound, sight and touch. It suits brands oriented towards the senses, such as food and drink.

It gives a richer understanding of the relationship between consumer and brand, and makes it easier for respondents to discuss something that on the surface seems ordinary.

It suits sensory categories such as food and beverage.

Multivariate analysis means analysing many variables together in order to understand a population more deeply than any single variable allows.

The most common forms include cluster analysis, which allocates individuals into groups based on how similar they are across several characteristics, and factor analysis, which reduces many correlated measures to a smaller number of underlying dimensions.

It underpins customer segmentation research.

Mystery shopping is an approach in which researchers pose as customers and carry out predefined tasks in order to assess the experience a brand is delivering.

In a retail setting the shopper may assess tidiness and cleanliness, signposting and ease of finding items, displays and merchandising, ease of checkout, and the willingness and availability of staff.

It is conducted across a brand’s own outlets and those of competitors, and is used widely by automotive, travel, hotel, banking and food businesses, and by public-facing government departments.

Brandspeak uses it within customer experience research.

Net promoter score is calculated from a single question asking how likely someone is to recommend a brand, on a scale of nought to ten.

Those scoring nine or ten are promoters, seven and eight are passives, and nought to six are detractors. The proportion of detractors is subtracted from the proportion of promoters, giving a score between minus 100 and plus 100.

It is widely used and easy to compare across time and organisations. Its limitation is that it reports a number without explaining it, which is why it is normally accompanied by a follow-up question asking why.

Brandspeak reports NPS alongside diagnostics in customer satisfaction research.

Neuromarketing applies techniques drawn from neuroscience and psychophysiology to measure responses to marketing stimulus that people cannot easily report themselves.

Technologies include functional magnetic resonance imaging and electroencephalography, which measure activity in the brain, and sensors that detect changes in breathing, heart rate and galvanic skin response. Eye-tracking records where attention falls, and facial coding classifies expressions into inferred emotions.

A good deal of marketing behaviour is shaped by factors outside conscious awareness, which is the argument for measuring response directly rather than relying on what people say. How far each individual technique delivers on that promise varies considerably, and claims in this field should be read with care.

See neuromarketing research services.

New product development research supports the development of new products and services. The term covers both qualitative and quantitative approaches.

It may be used to identify and evaluate commercial opportunities, assess early-stage propositions, or let respondents interact with developed ideas, prototypes and samples.

Brandspeak offers this as innovation and NPD research.

Non-probability sampling is any approach in which members of the population do not have a known chance of selection. Quota, convenience and snowball sampling all fall into this category.

Most commercial research is non-probability, which is why findings are properly reported as indicative of the sample rather than as statistically projectable to the population.

See quantitative research.

Non-response bias occurs when those who take part differ systematically from those who do not, so the sample no longer reflects the population.

It is a greater risk than a low response rate on its own, because the rate matters only if the people missing are different from those who answered.

It is a design consideration in all quantitative research.

An omnibus is a quantitative survey on which several clients buy space, each adding their own questions. The survey goes to a pre-recruited, usually nationally representative sample, often of around 1,000 respondents.

Omnibuses typically run weekly, so answers come back quickly. Costs are shared between the participating brands, which makes the approach economical where only a few questions are needed.

Results are usually cross-referenced against classification data and supplied as data tables or a short report.

See quantitative research.

An online bulletin board is a form of qualitative research run over several days, in which a moderated group communicates through a message board.

Respondents take part in discussions, pop-up surveys, reactions to stimulus and uploads of their own material. They log on at times that suit them, which keeps drop-out low, and the moderator can change focus or stimulus as the research develops.

It is one of Brandspeak’s qualitative research methods.

An online focus group is a discussion of 60 to 90 minutes between a moderator and usually eight to ten respondents, conducted over the internet.

They may be run with webcams, so that moderator and respondents can see one another, or as text only. They offer savings in time and cost and make it possible to bring together respondents who are geographically dispersed.

See focus groups and depth interviews.

Online market research is a catch-all term for any research conducted over the internet, covering qualitative and quantitative approaches and studies lasting from a few minutes to several months.

See Brandspeak’s market research services.

Online qualitative research covers any qualitative method conducted online, including online focus groups, bulletin boards, research communities and depth interviews.

See qualitative research.

Order effects are the influence that the sequence of questions or stimulus has on the answers given.

An item shown first is often rated more favourably, and an earlier question can frame a later one. Rotating the order between respondents is the standard control, and a design that does not rotate should be treated with suspicion.

Rotation is built into Brandspeak’s quantitative research designs.

Packaging testing evaluates whether a pack is noticed on shelf, understood at a glance and attributed to the right brand, and whether it conveys what it is intended to convey.

It combines measures of visual attention with questions on comprehension, appeal and fit with the brand. It is also used to assess the risk in a redesign, where the danger is losing recognition built up over years.

Brandspeak offers this as packaging research.

A paired depth interview, or paired depth, is conducted with one moderator and two respondents. It is also called a dyad.

It allows the researcher to probe each person’s experience while using similarities and differences between them to explore the subject more deeply. The pair may be in a relationship and have made a purchase decision jointly.

A friendship pair is a variant in which two people who know each other are interviewed together, which can help where the subject is sensitive.

See focus groups and depth interviews.

A persona is a composite description of a target customer, built from research and used to keep teams aligned on who they are designing for.

A persona drawn from segmentation data is a planning tool. One invented in a workshop is a shared assumption with a photograph attached, and it should not be mistaken for evidence.

Personas built from data come out of customer segmentation research.

Personification is a projective technique in which respondents describe a brand as a living entity, usually a person or an animal.

Where a person, respondents may be asked to describe their personality, interests, friends and what motivates them, or to imagine how they would behave in a particular situation. It unlocks impressions, vocabulary and imagery that direct questioning does not reach.

The related term anthropomorphism describes the same impulse to assign human characteristics to things that do not have them.

It is a qualitative research technique.

A pilot study is a small study run in advance of the main exercise, to test key elements of the approach before full roll-out.

Pilots are most common before quantitative studies, because a questionnaire is difficult to change once fieldwork has started. In qualitative research, most aspects of the approach can still be adjusted after the project begins.

See quantitative research.

A positioning statement encapsulates the place a brand intends to occupy in the target customer’s mind. Internally it guides planning, advertising, goal setting and measurement; externally it is expressed in condensed form through communications.

It sets out the brand’s key benefits within the context of the competition, the target audience and the reasons to believe. A common template runs: [brand] is the only [frame of reference] that delivers [benefit] to [customer] who want [need state], because it [reasons to believe].

Positioning is developed and tested through brand development research.

Pre-placement is the practice of placing stimulus material with respondents in advance of the research, usually some days before, so that they can give informed feedback rather than a first impression.

It is used in qualitative research.

A pre-task is an activity given to respondents before research begins, such as studying stimulus or keeping a diary of relevant daily activity.

Its purpose is to have respondents thinking about the subject, and about their own behaviour in relation to it, before the session starts, which makes the discussion more focused and more specific.

It is standard practice in Brandspeak’s qualitative research.

Pre-testing evaluates highly finished advertising material before launch. It may be qualitative or quantitative.

Its purpose is to confirm that a concept which performed well in earlier research has translated successfully into the finished execution, and that nothing in production is obscuring clarity or impact.

See advertising development research.

Pricing research establishes what customers will pay, how demand changes as price changes, and how a price should be structured across a range.

It covers willingness to pay, optimal price points, price elasticity and price architecture. The main techniques are Van Westendorp, Gabor-Granger, brand price trade-off and conjoint analysis, each suited to a different question.

Brandspeak delivers pricing research including conjoint analysis and Van Westendorp.

Primary research is conducted directly with the subjects of the study, rather than drawing on studies that already exist, which is secondary research.

The term covers both qualitative and quantitative approaches, including focus groups, surveys, field tests and observation.

It is distinguished from secondary research.

Probability sampling is any approach in which every member of the population has a known, non-zero chance of selection.

It is what allows results to be projected to the population with a calculable margin of error. It is rarer in commercial research than the language of reporting often implies.

See quantitative research.

Product positioning research either identifies the optimal territory for a positioning, after which options are created and tested, or tests and strengthens a positioning that already exists.

See brand development research.

Product testing evaluates a physical product with its intended users, measuring appeal, performance and purchase intent.

It may be monadic, with each person testing one product, or comparative. Testing in the home over time usually gives a more reliable answer than a single trial at a central venue, because performance emerges with repeated use.

Brandspeak runs product testing within innovation and NPD research.

Projective techniques are used where respondents would otherwise struggle to articulate their impressions of a brand, or would feel inhibited in doing so.

Respondents may be asked to imagine a brand as a vehicle, a tree or an animal, to write its obituary, or to describe the party it would host, including the location, mood, music and guests. The indirection is the point: it bypasses the rationalising that direct questions invite.

Projectives are used throughout Brandspeak’s qualitative research.

Psychographics group people by shared psychological characteristics such as attitudes, values, outlook and concerns, rather than by demographics.

People of different ages, social grades or life stages may share an adventurous disposition that leads them towards unusual holidays and unfamiliar food. Where such characteristics can be identified reliably, they form a sound basis for segmentation and targeting.

Psychographic segments are built through customer segmentation research.

The purchase funnel describes the stages a buyer moves through, typically from awareness to consideration, preference, purchase and repeat.

It is a useful reporting structure and a poor description of how people actually behave, since real journeys loop, skip stages and often begin again from the middle.

Funnel measures are tracked within GrowthTrack.

Qualitative research is the non-numerical exploration of a subject in order to uncover perceptions, feelings, motivations and opinions. It answers why rather than how many.

It is conducted through focus groups and depth interviews, face to face or online, and through bulletin boards and research communities. It uses discussion guides that respond to the flow of conversation and may be adjusted as the project runs.

The interpretative judgement of the researcher is central to turning qualitative findings into usable insight.

Brandspeak’s qualitative research is director-led.

Quantitative research covers approaches that measure incidence and the statistical relationships between variables, reflecting perceptions, needs and behaviours. It answers how many and how much.

It uses structured questionnaires with set questions and limited responsiveness to context. It is most often conducted online, but may also be face to face, by telephone or by kiosk. Sample sizes are larger and outputs are subject to statistical analysis.

See quantitative research.

Quota sampling sets targets for the number of respondents with particular characteristics, usually age, gender and region, and fills them as fieldwork proceeds.

It matches the sample to the population on the characteristics used to set the quotas. It does not guarantee representativeness on anything else, because respondents are not selected at random, which makes it a non-probability method.

See quantitative research.

Random sampling is the simplest form of probability sampling, in which every person has the same chance of being chosen.

Random selection removes systematic bias from how the sample is chosen. It does not remove sampling error, and the result will still differ from the population by a margin the sample size determines.

To be accurate it relies on having a complete list of the population, which in commercial research is rarely available.

See quantitative research.

Reciprocity bias is the impulse to repay what we have received. When people are treated generously they tend to respond in kind, and when treated badly they tend to respond in kind to that as well.

It is widely used in sales and marketing. A free sample or trial creates a felt obligation, which raises the likelihood of a purchase or of agreement to a request that follows.

It is one of the biases considered in brand strategy consulting.

Regression analysis compares the relationship between dependent and independent variables. It shows how the typical value of the dependent variable changes when one independent variable changes while the others are held constant.

It is widely used in prediction and forecasting, and it underpins driver analysis in research.

It underpins the driver modelling in GrowthTrack.

A representative sample is a subset of a larger population that reflects the characteristics of that population.

It is used where gathering data from everyone is impractical. The sample should reflect the profile of the wider group across all relevant characteristics, which may include age, gender, region, working status and income.

See quantitative research.

Research design is the specification of the research to be undertaken. It will cover methodology, sample, geographic focus, data collection approach, analysis and the format of reporting, and it differs according to whether the work is qualitative or quantitative.

Brandspeak designs both qualitative and quantitative studies.

Research method refers to the approach chosen to deliver the required insight. Examples include focus groups, depth interviews, bulletin boards, ethnography and surveys.

Brandspeak’s methods are set out under research methods.

A research panel is a group of people recruited to take part in surveys over an extended period. Members are incentivised, usually with points redeemable for cash or goods.

Panels are mainly quantitative. Members may be invited into longitudinal studies repeated over time, or into one-off studies on a different subject each time.

Most panels are consumer-based, but they can also cover professional audiences such as general practitioners, dentists or IT decision makers.

Panels supply the sample for most quantitative research.

In survey research, the response rate is the proportion of those invited who complete the survey.

A low response rate raises the risk of non-response bias, because those who answer may differ systematically from those who do not. In direct marketing the term is used differently, to mean responses as a percentage of those contacted.

See quantitative research.

The sample is the group of individuals being researched. Sampling is the process of selecting them.

A sample is configured to represent a larger population of interest, so that results based on it can be read as indicative of that population.

See quantitative research.

A sample frame is the list from which a sample is drawn, such as a customer database or a research panel.

Any group absent from the frame cannot appear in the sample, however well the sampling is executed. This makes the frame a common and under-examined source of bias.

See quantitative research.

Sample size is the number of respondents in a sample. It determines, in part, the reliability of the results, and it governs how finely the data can be read.

The more subgroups, markets or segments that need to be reported separately, the larger the total sample has to be, because each needs a base of its own.

See quantitative research.

Sampling error is the difference between a result from a sample and the true figure in the population, arising purely from having sampled rather than counted everyone.

Unlike bias it is random, it can be quantified, and it reduces as sample size grows. It cannot be removed, only estimated and allowed for.

See quantitative research.

A screener is a short set of filter questions used to establish whether someone is eligible to take part in a study. The questions reflect the recruitment criteria agreed between client and agency.

The screener also determines the incidence rate, and with it a significant part of the cost of fieldwork.

Screening criteria are agreed at the design stage of any market research project.

Secondary research uses data that already exists, rather than commissioning new research into the subject. It includes published reports, official statistics, trade sources and the client’s own records.

Brandspeak offers this as secondary research.

Selection bias arises when the way participants are chosen makes the sample unrepresentative in a systematic way.

A satisfaction survey sent only to active customers is a common example, since those who have left are the people most likely to have something useful to say.

It is a design consideration in customer satisfaction research in particular.

Self-ethnography explores behaviour, decisions and opinions through the eyes of the subject, who records their own experience rather than being observed by a researcher.

Participants respond to prompts from a moderator, recording thoughts and decisions through diary entries, photographs and video.

See auto-accompanied ethnography.

Semiotics is the study of signs and of how meaning is made. The word comes from the Greek for sign.

People use semiotic systems daily: the words chosen to express something, the clothes selected in the morning, the food judged right for a moment. Research uses semiotics to understand how consumers decode their environment and ascribe meaning to brands.

That understanding helps in creating communications that work at both conscious and subconscious levels.

It informs Brandspeak’s approach to communications research.

Sentiment analysis examines large volumes of online comment, about a brand, a person or a campaign, in order to determine the underlying sentiment as positive, negative or neutral. It is a central part of social listening.

It supports competitor analysis and brand monitoring.

In a sequential monadic design each respondent evaluates several items one after another, each in isolation rather than in direct comparison.

The order is rotated between respondents to control order effects. It is more economical than a pure monadic design when several items must be compared, at the cost of some carry-over between evaluations.

Brandspeak has written on monadic versus sequential monadic testing.

Share of search is a brand’s share of total search volume within its category.

It is inexpensive to track and has been shown to correlate with market share, often moving ahead of it. That makes it a useful early indicator alongside survey-based measures rather than a replacement for them.

It complements survey measures in brand tracking.

Share of voice is a brand’s share of total category advertising within a period.

Compared against market share it indicates whether a brand is spending above or below its weight, which is one of the longer-standing predictors of whether a brand will grow or decline.

It is read alongside brand tracking measures.

A shopper intercept is conducted with customers immediately after an experience of the brand, whether they purchased, intended to purchase, or left without buying.

Consumers are approached during or just after their visit and screened before a short interview. Intercepts are used to evaluate the experience just received, to capture reactions to sampling, to identify which brands were under consideration, and to establish why no purchase was made.

They are used in retail and experiential marketing research.

Snowball sampling recruits participants through referrals from those already taking part.

It is used where a population is small, hard to identify or reluctant to come forward. It produces a sample clustered around existing networks, which limits what can be claimed from it.

It is sometimes necessary in B2B research with very niche audiences.

Social desirability bias is the tendency to give answers that show the respondent in a good light.

It inflates reported exercise, saving, recycling and charitable giving. It is reduced by self-completion formats, by indirect questioning, and by asking about behaviour rather than intention.

Mitigating it is part of questionnaire design in quantitative research.

Social media listening monitors social channels for mentions of a brand, product, service or competitor, and analyses what is found in order to identify actionable insight and emerging themes.

It supports competitor analysis.

Social monitoring analyses social media for mentions of brands, individuals or campaigns. It differs from social listening in being more passive, focusing on what has already happened rather than on what the pattern suggests may happen next.

See competitor analysis.

Staff research is conducted among the employees of an organisation. The opinions, experience and ideas of employees are often a valuable and under-used source of insight, and they bear directly on the experience customers receive.

It is closely related to customer experience research, since employee and customer experience move together.

Statistical significance is the judgement that an observed difference is unlikely to have arisen by chance, given the size of the sample.

It says nothing about whether the difference is large enough to matter commercially. A difference can be statistically significant and strategically irrelevant, and on a large sample it frequently is.

Significance testing is applied throughout quantitative research.

Strategic development research is qualitative work undertaken to identify the insight needed to shape strategy, usually for a new brand or a marketing campaign.

It addresses the strategic building blocks in detail: the target customer, their needs, expectations, behaviours and points of reference; and competitor brands, both in isolation and relative to the client’s own.

See brand strategy consulting.

Stratified sampling divides a population into groups, or strata, and samples within each, so that every group is represented in known proportion.

It gives more reliable results than simple random sampling when the groups differ from one another, and it allows small but important groups to be reported separately by setting a minimum base for each.

See quantitative research.

A sub-brand sits beneath a parent brand, carrying its own name and character while drawing on the parent’s credibility.

Research on sub-brands usually turns on how visible the parent should be, and on whether the sub-brand adds clarity for the buyer or simply adds another name to remember.

Architecture questions are addressed through brand strategy consulting.

System 1 and System 2 describe two modes of thought. System 1 is fast, intuitive, emotional and effortless. System 2 is slow, deliberative and rational, and it requires attention.

Most everyday choices are made through System 1, with System 2 engaged for decisions that carry greater consequence. The distinction matters commercially because a great deal of brand choice happens quickly, with emotional cues taken from visual elements such as packaging and merchandising rather than from deliberate comparison.

The framework was popularised by Daniel Kahneman in Thinking, Fast and Slow.

It informs how stimulus is designed in communications research.

Tabulation is the arrangement of data into tables for analysis. Data tables are the standard deliverable from a quantitative study alongside the report.

Data tables are a standard deliverable in quantitative research.

A telephone interview is a research interview conducted by phone, usually from a central unit running several surveys at once. It remains the most practical method for many B2B audiences, who are easier to reach by telephone than online.

Brandspeak uses telephone interviewing particularly for B2B research.

Text analytics examines unstructured text written by or about customers, in order to identify sentiment, issues and opportunities that other methods would not surface.

It is performed by software using text mining and natural language processing, and it makes large volumes of open-ended response usable.

It is used to code open-ended responses in quantitative research.

A touchpoint is any point of contact between a customer and a brand, from an advertisement to a delivery note to a conversation with support.

Experience research maps them because the impression a brand leaves is formed across all of them, and rarely at the point the brand itself considers most important.

Touchpoints are mapped in customer experience research.

Trend analysis examines how a measure has moved over time in order to separate a genuine direction of travel from normal variation.

In tracking it is the discipline that prevents a single wave being over-read. Two points make a line; three or more make a trend.

It is the discipline behind reading brand tracking results.

A triad is a group discussion involving three respondents. It sits between a depth interview and a full focus group, offering some group interaction while allowing each participant more time.

See focus groups and depth interviews.

TURF stands for total unduplicated reach and frequency. It is a form of analysis used to estimate the combined reach of a set of options and the overlap between them.

It identifies which combination reaches the largest number of people, which makes it useful in deciding product ranges, message sets and media placement where budget is limited.

It is used to optimise message and range decisions in quantitative research.

One TVR, or television rating, equals 1% of a defined target audience viewing an advertisement at least once. Campaign TVRs are the sum across all spots, which is equivalent to reach multiplied by average frequency.

See advertising development research.

Usability testing evaluates a product or service, typically a website, device or application, with a group of target users performing defined tasks.

Its purpose is to establish whether the product is fit for purpose and to identify the specific improvements needed to deliver a satisfactory experience and uptake.

A well-run usability test identifies essential improvements and also informs brand strategy, user guides and supporting materials. Tests are often conducted in facilities configured so that observers can watch unseen.

Brandspeak runs usability work within customer experience research.

A usage and attitude survey establishes consumer awareness of a product or service alongside related attitudes, tastes and behaviours. It may be conducted once or repeated so that change can be tracked.

U&A surveys inform brand and product development, targeting and communications, and they are often the first substantial study commissioned in an unfamiliar market.

U&A studies sit within quantitative research.

A user journey is the route a person takes online in order to accomplish a goal. User journey mapping records that route and its pain points so that designers can improve the experience.

See customer experience research.

A value proposition communicates a brand’s value in terms of its most compelling features and benefits, including any unique selling points.

A strong proposition contains both functional and emotional benefits. A car brand may offer more safety features as standard than any rival, which is functional; the corresponding emotional benefit is peace of mind. Uber’s proposition, the smartest way to get around, encapsulates several practical benefits in a single line.

The proposition works in conjunction with the brand’s positioning.

Propositions are developed and tested through brand development research.

The Van Westendorp price sensitivity meter establishes an acceptable price range using four questions: at what price the product would be too expensive to consider, too cheap to be trusted, starting to feel expensive, and a bargain.

Responses are plotted to identify the range within which price is acceptable and the point of least resistance. The approach works best in established categories where respondents have a frame of reference.

Brandspeak delivers Van Westendorp within quantitative research.

A variable is a quantity or property with an assigned value that may differ between respondents or change over the course of the research.

See quantitative research.

In research, video streaming most often refers to live streaming qualitative sessions over the internet, so that client viewers unable to attend in person can watch from elsewhere.

It is offered on Brandspeak’s qualitative research projects.

A viewing facility is a venue built for research, comprising back-to-back rooms separated by a one-way mirror. Clients observe proceedings between moderator and respondents from the adjoining room.

Most client rooms also offer monitors, live feeds and streaming for those unable to attend.

See focus groups and depth interviews.

In research, virtual reality allows respondents to experience stimulus in three dimensions without the expense of building prototypes.

Its most useful applications are in evaluating in-store environments, packaging and merchandising, where the alternative is either a flat image or a costly physical mock-up.

Its clearest application is in packaging research.

Voice of the customer describes a programme that gathers customer feedback continuously across channels and feeds it into decisions.

It differs from a survey in being ongoing and in combining sources, including unsolicited feedback such as complaints and reviews alongside questions that are actively asked.

See customer experience research.

Wearables are items of clothing or accessories with embedded technology capable of collecting and transmitting data about the wearer.

Devices such as wristbands record physical activity, sleep quality and heart rate. In research their value lies in capturing behaviour passively, without relying on recall, particularly in health and wellbeing categories.

They are used within ethnographic research.

Weighting increases or decreases the statistical weight given to the responses of a particular group, so that the group is represented in proportion to its share of the population.

A group that is over-represented in the achieved sample can be weighted down so that it does not distort the findings, and an under-represented group weighted up so that its views carry their proper share.

Heavy weighting is a warning sign rather than a solution: if a group has been weighted up substantially, the findings for that group rest on a small number of actual respondents.

See quantitative research.

Willingness to pay is the most a customer would pay for a product before choosing not to buy it.

It is estimated rather than observed, through techniques such as Van Westendorp, Gabor-Granger and conjoint analysis. Stated willingness generally runs above what people do when the money is real.

Pricing techniques sit within quantitative research.