Find out what drives your customers' choices.

Pricing, product, segmentation and loyalty questions answered with choice modeling, segmentation and driver analysis. Results come with the tools to keep using them, and with methods that hold up when finance, product and leadership ask how the answer was reached.

Conjoint and discrete choice, MaxDiff, segmentation, key driver analysis, pricing models, TURF, machine learning.

Applications

  • Brand research

    How customers see your brand next to competitors, and which perceptions drive choice. Awareness, usage and image are tracked over time, and driver analysis shows where to focus.

  • Pricing research

    How demand and revenue change with price, and where the best price point sits. Conjoint, Van Westendorp and Gabor-Granger models produce price curves and revenue estimates, with a simulator for testing price points against competitors.

  • Product and feature decisions

    Which features customers would trade for which, and which configurations win share. Respondents choose between realistic products, so results reflect choices rather than stated preferences.

  • Concept and message testing

    Which concepts, claims or ads to take forward, and why. Monadic and sequential designs, MaxDiff and driver analysis identify the strongest options and what makes them work.

  • Advertising and campaign effectiveness

    Whether a campaign changed awareness, perception or behavior, and by how much. Pre/post and test/control designs, with weighting and modeling to isolate the effect.

  • Customer segments

    Groups defined by attitudes, needs and behavior, not demographics alone. Each segment is profiled and sized, and a short typing tool assigns new respondents or customer records to the right group.

  • Satisfaction and loyalty

    Which attributes actually move satisfaction, loyalty or purchase intent, and which only look important. Driver models are matched to the data, including relative importance methods when attributes overlap.

  • Product lines and message sets

    The smallest set of products, flavors or messages that reaches the most people. MaxDiff ranks the options and TURF finds the combination.

  • Category structure and language

    How customers group products, needs or terms, and what they call them. Card sorts and ranking exercises, analyzed with clustering and multidimensional scaling, shape categories, navigation and naming.

  • Tracking and KPIs

    Measures that stay comparable as questionnaires, samples and populations change. Weighting, derived variables and codebook alignment keep trends consistent across waves.

Methods

The method follows the question.

Conjoint and discrete choice
Which features and prices customers will choose. Hierarchical Bayes and latent class estimation.
MaxDiff
What matters most, in order, with individual-level scores for features, messages or claims.
Segmentation
Which groups of customers to treat differently. K-means, hierarchical and latent class methods.
Key driver analysis
What actually moves satisfaction, loyalty or intent.
Pricing models
The right price without losing customers.
TURF
The product line or message set that reaches the most people with the fewest items.
Factor and path analysis
The underlying dimensions in a long attribute list, and how perceptions lead to outcomes.
Machine learning
Classification and prediction with models built to be explained as well as accurate.
Weighting
Results that reflect the real population, with design effects reported.

Services

  • Study and experimental designChoice task design, attribute structures, sample size and the analysis plan.
  • Analysis of existing dataSurvey, customer and transaction data you already hold.
  • Market simulatorsSpreadsheet or web tools for testing products, prices and competitive scenarios.
  • Typing toolsShort classifiers delivered as specifications, spreadsheets or code for survey platforms and client systems.
  • ReportingFindings and recommendations in plain language, as reports, presentations or data files.
  • Data collectionFielding through Research Services when new data is needed.

Approach

  1. Start from the decision

    A price, a product line, a target market. The decision sets what has to be measured and how precise the answer needs to be.

  2. Design and estimate

    The method, sample and questionnaire are designed together, so the data supports the model. The analysts who design the study run the estimation.

  3. Deliver results and tools

    Findings and recommendations, plus simulators and typing tools that keep the results in use after the report.

Past performance

Drive-Thru Menu Board Conjoint Study - Major Beverage Supplier

A major beverage supplier needed to measure the patron experience of ordering from digital menu boards typical of fast food restaurant counters and drive-thrus. The study required an advanced conjoint design to be administered in a simulated in-person ordering environment.

Survey Launch devised a methodology and supporting technical application that broadcast high-resolution dynamic menu content on flat-screen LCD monitors in a mock ordering area, synchronized with a discrete choice exercise running on the interviewer's laptop. Interviewing areas were set up at central location testing facilities simulating both restaurant counter and drive-thru ordering environments. Survey Launch designed the methodology, built the technical application, set up the facilities, and trained interviewing staff.

Capabilities demonstrated: Advanced conjoint design, in-person intercept research, custom technical development, CLT facility management, QSR sector experience.

Key Performance Indicator Tracking - University System

A university conducts periodic surveys among prospective students, current students, and alumni to track attitudes, behaviors, outcomes, and demographics across cohorts over time. The study varies across waves based on changes in program initiatives and objectives.

Survey Launch performs a range of analytical work across these data including data transformations, weighting, recalculation of derived variables across multiple segments using disjoint time-series data, cluster analyses, and key driver analysis to identify changes and update KPIs. The firm also advises on variable development, transition strategies, and codebook alignment across waves.

Capabilities demonstrated: Longitudinal KPI tracking, advanced analytics, derived variable development, multi-cohort comparative analysis, higher education sector experience.

Statewide Public Input Study - State Early Childhood Agency

A state early childhood agency was required to incorporate public input into a federally mandated plan and needed a systematic method for collecting, analyzing, and documenting community feedback at scale.

Survey Launch designed and administered a public input process combining in-person sessions, virtual meetings, and a dedicated online portal. The most recent cycle collected 2,088 open-ended comments across twelve survey questions. Analysts conducted systematic reviews of all responses, developed a comprehensive coding framework, and transformed unstructured qualitative feedback into structured quantitative data suitable for statistical analysis.

Deliverables included an executive summary, coded data tables for each question, and complete verbatim responses organized by question. The analysis enabled the agency to document community priorities clearly and incorporate them into the state's federally required plan.

Capabilities demonstrated: Public engagement research, qualitative coding and quantitative transformation, state agency research, federally mandated evaluation requirements.

Questions

Can you analyze data we already have?

Yes. Existing survey, customer and transaction data can often support segmentation, driver analysis or predictive models. A first look shows what it can support.

How is the right method chosen?

Together, and ideally early. Survey Launch joins discovery alongside your stakeholders, who know the business best, then recommends the approach that fits the decision and explains why.

What is your analysis workflow?

Analysts follow established processes for cleaning, exploring and modeling the data. All work is done in version-controlled code, so every result can be reproduced and audited.

How large does a study need to be?

It depends on the method and how many groups need to be compared. The design step sets the number before anything is fielded.

What decision are you facing?

Describe it in a few sentences. A senior analyst will suggest how to answer it.

Describe your decision