Statistical analysis & modelling

At DJS Research, we pride ourselves in going far beyond just number-crunching. We use advanced statistical techniques to unlock powerful insights, harnessing the full potential of data and helping our clients make confident, evidence-based decisions. We specialise in making complex analysis intuitive and actionable. Whether it’s understanding consumer preferences or forecasting market demand, our techniques are designed to maximise clarity, relevance, and commercial impact.

Some of the core techniques we specialise in include:

• Predictive modelling: build forecasting models that anticipate future behaviours like brand switching, purchase intent, or churn. Using techniques like linear and logistic regression, we help clients act before opportunities – or problems – arise. 
• MaxDiff: We use this powerful survey and analysis technique to help identify what people value most by asking them to choose the best or worst options from a set. This is ideal for identifying which features or messages from a ‘longlist’ matter most to consumers or business decision makers – it forces clear trade-offs and reveals relative preferences with greater precision than simple appeal ratings. 
• Conjoint analysis: This technique uncovers how people make trade-offs between product features when making decisions. It’s essential for designing products that reflect what people truly value, and is often a next step after using a MaxDiff approach. 
• Segmentation: We group consumers into distinct, targetable segments based on shared characteristics like attitudes or behaviours. This enables precise targeting, tailored messaging and strategic planning aligned to the needs of different market groups. Statistical techniques like correlation analysis and regression are often used to identify which variables most strongly influence segment formation.
• Price sensitivity modelling: We model how pricing changes affect demand and potential revenue. Combined with market size estimation, this provides a full picture of commercial opportunity. We use a range of techniques to explore pricing, including Gabor-Granger and Contingent Valuation.
• A/B testing: Testing different variations of a message, campaign or product enables us to identify which performs best. We apply rigorous experimental design methods to ensure results are robust, reliable and replicable.
• TURF analysis: Short for Total Unduplicated Reach and Frequency, TURF identifies the optimal set of products, features or messages that appeal to the widest, most distinct audience – ideal for portfolio optimisation and new product development.

While these techniques might sound sophisticated or academic, they’re by no means a ‘dark art’ and we never hide them away behind closed doors. We’re committed to making our work clear, accessible, and genuinely useful – both for our immediate clients and their wider stakeholder teams.