L4

Sony BetaMax (Cooper, 2000): what does it teach about market research?

Morita (Sony co-founder) rejected market research for new products ("we are the experts"); Sony missed demand for full-length film rentals (60-min tapes, incompatible format) and copyright conflict with Universal; lesson from Souchon: always do some research, you don't know best.

Why is researching the international marketing environment important? (7 reasons)

Informs most IM decisions: 1) anticipate need to internationalise (competitor analysis); 2) segment/target markets (country screening); 3) set measurable objectives (market quantification); 4) manage foreign partnerships (reward systems, negotiations); 5) design the mix (product adaptation); 6) control success (research designs); 7) adapt theory (service quality measurement).

Market research method options

Primary (first-hand, collected for this problem) or secondary (already published for another purpose); each can be quantitative (counted/measured) or qualitative (meanings, experiences, why); plus big data as an added source.

Evaluate secondary data for international markets

For: cheap, fast, large-scale, benchmarking/trends, informs primary research, gives census/national stats/industry reports. Against: hard to compare across countries, quickly outdated, not always available, information overload, limited relevance, no control over quality/bias. Best used first to maximise efficiency, not alone.

Primary research design: 4 steps with the lecture's two examples

1) Problem/objectives; 2) collection method; 3) sampling and frame; 4) analysis. Market sizing of non-woven hi-vis jackets in Spain = quantitative survey, random sample, regression. NPD for Scottish raspberries in France = qualitative focus group, purposive sample, content analysis.

Qualitative methods: collection and analysis (Hillside example)

Collection: in-depth/focus group interviews, diaries, archival, netnography, videography, ethnography, observation. Analysis: discourse (language, e.g. branding), narrative (customer journeys), content (themes, e.g. eWOM), thematic coding, NVivo, within/cross-case networks and matrices. Hillside: netnography of customer reviews analysed by discourse analysis to inform branding.

Quantitative data collection and the main challenge

Structured questionnaire via mail, telephone (interviewer bias: more favourable answers), mobile, internet/email or personal interview. Main challenge is avoiding bias: representative sample and careful question formulation.

Four rules for survey questions (with the lecture's bad examples)

Neutral (not "Do you care enough about your family to…insurance"); unambiguous (not the double negative "prefer not to travel in a non-smoking flight"); answerable (not "least memorable experience"); easily understood (not jargon like "arbitrary surcharges for ITX packages").

Three descriptive analyses (lecture example, N=81, 1–7 value-for-money scale)

Frequencies = how many? (e.g. 20 rated it 4); means/SDs = on average? (mean 3.99, σ 1.59); cross-tabulations = how many of each? (younger groups rated higher; 55+ mostly rated 1–2).

Standard deviation formula (as in lecture)

σ = √[Σ(xi − μ)² / N]; xi = each value, μ = population mean, N = number of values, Σ = sum; shows how spread out ratings are around the mean; a low σ means respondents largely agree.

Two inferential analyses with the lecture's results

Comparisons (segment A vs B): independent t-test, British mean 4.42 vs French 3.29, t = 3.31, sig .001, so British rate it significantly higher. Relationships (more x, more y?): Pearson r = −.907, sig < .001, so older consumers rate it much lower.

Multivariate and structural inferential analysis: lecture example and link to Topic 1

Regression: Export performance = β0 + β1(planning) + β2(improvisation) + ε; SEM tests NPS as mediator between planning/improvisation and growth/profits (γ and β paths); judge each effect by its p-value. Link: Topic 1 export decision-making (Nemkova et al., 2015 on improvisation vs planning).

What research design does Souchon recommend and why?

Method depends on objectives; methods are complementary so use more than one: secondary data first for efficiency, then qualitative (e.g. in-depth interviews) to gain insight, then quantitative (e.g. survey) to predict and quantify.

Why is international primary research difficult?

Multi-country, multi-cultural and multi-linguistic settings make it expensive to collect, difficult to coordinate and difficult to establish comparability; sampling frames can be hard to find in some countries.

Cross-national equivalence: definition and three families (Carter, 2019)

Equivalence = research measures the same thing in the same way in every country, so differences are real not artefacts (bias). Families: construct (same meaning), measurement (same measuring), sampling (comparable people); Carter sets out 11 untested propositions (P3–P11 cover equivalence).

Construct equivalence and its sub-types (Carter, 2019)

Does the construct mean the same and matter equally across cultures? Conceptual (operational P3, e.g. "eating out" = restaurant vs street stall; item P4); functional P5 (bicycle = leisure in UK, transport in NL/China); category P6 (beer as alcoholic vs soft drink).

Four types of measurement equivalence (sample exam question)

Translation P8: same meaning after translating ("extremely agree" vs "completely agree" endorsed differently in US/UK/Canada/France); fix = back-translation. Calibration P7: same units (£5 vs $5, kg vs lb; exchange rates). Metric P10: same psychometric structure, scale intervals and origin. Score P9: does 4/5 mean the same?

Score equivalence: response styles and fixes

Extreme responding (US, Latin America pick 1 or 5), middle responding (Japan, China pick 3), acquiescence (agreeing to be polite); so Brazil's 4/5 ≠ Japan's 4/5. Fixes (Carter): standardise scores, confirmatory factor analysis; further reading De Langhe et al. (2011).

Sampling equivalence and Carter's two-stage segmentation

Sampling P11: avoid convenience/student samples; use matched, proportionate or weighted samples. Two-stage: 1) cluster countries on macro economic/cultural data (identifiable, accessible, stable); 2) segment consumers within clusters on benefits/attitudes/behaviour (responsive, actionable); P1/P2 claim higher profits than either alone.

Exam prompt: "Outline the four types of measurement equivalence and evaluate how they should inform your international research design and/or analysis" (answer skeleton)

Intro: equivalence issues are unique to IM research and bias decisions (Carter, 2019). Each type: define, give lecture plus own example, design fix (back-translation, local units, pre-tested scales, standardising/CFA). Evaluate: costly and slow; fixes not detailed and propositions untested (Carter limitations); wider reading (Pourmasoudi et al., 2024; De Langhe et al., 2011). Conclude: build checks in at design stage.