Skip to main content

Why do we do statistical analysis?

By 10 May 2022February 22nd, 2023Blog, Insights

Humans don’t behave in a simple, logical manner – they don’t act as they should. Retirees go to nightclubs, teenagers play bingo, they don’t conform to stereotypes. We need methods to understand behaviour based on who your customers really are and how they behave, not how old they are, or where they live.

At Launchpad Research we use statistical analysis techniques like cluster analysis to create actionable segmentations of your customers based on their attitudes and behaviour, not stereotypes. This allows us to create segmentations that have actionable insight that will allow your business to truly understand your customers and plan for success.

Another issue with humans not behaving as we would like is seen in the way they respond to questionnaires – what they say they do is not always what they do. To resolve this, we use statistical techniques such as Key Driver Analysis and Bayesian Network Analysis to see what is truly driving behaviour, be it recommendation, product purchase or satisfaction. Respondents indicate across a series of metrics and attitudinal statements which can be used to see what links to the desired outcome. This shows us what is important and what equally, what is not important, ensuring you spend your budget and time on what really matters to your customers.

Understanding human behaviour allows us to ask questions better too – using discrete choice techniques like MaxDiff allows us to understand relative and absolute attitudes towards attributes and brands in a way that the human brain can cope with. Ranking multiple brands on a page in a traditional manner is not something that is possible with accuracy at scale, swapping in a MaxDiff question allows for more accurate understanding.

Similarly, the best way to understand what your customer wants from your products or service is to use conjoint experiments to get them to highlight the importance of each individual element. This is done through the respondent selecting between complete products, not each element in isolation which they cannot do with accuracy. What is it that is making them choose you? Or your competitors? Which attributes are most important, and which make no impact? How does price interplay with this – can you increase the price without impacting customer choice? All these are better understood through conjoint experiments.

Most customers are interested in the same sets of products, so when developing new products how can you decide what products to introduce? Conducting TURF analysis helps us understand which additional products will bring in new customers, not just be bought by your existing set – the products that offer unique reach and frequency, not churn.

Get in touch to find out more about Advanced Statistical Analysis or any of the other ways in which we conduct quantitative research.