The quantitative research stage rounds off the four-pronged primary research programme of Brandheld.
The need to test our hypotheses from the previous stages might seem obvious.
Collecting and collating a large number of figures gives us the opportunity to statistically validate theories, market size, predict growth and compare behaviour or attitudes across different groups of people in different situations
But the best method to undertake this collection isn’t so obvious.
When designing the quantitative research, we made the following choices to ensure the data collected would be as accurate and as informative as possible:
Method of data collection: An online survey offers the most flexible approach to quantitative data collection. People can choose to participate at a time convenient to them, answering at their own pace without any biases caused by an interviewer (either directly through the way the questions are asked or the answers are recorded, or indirectly through a respondent feeling pressure to give a “socially acceptable” answer). However, online surveys contain other biases – the most serious being that the online population – and the online population taking surveys – are different to the population as a whole
Weighting factors: To correct the online population bias, we simultaneously asked several key questions in a face to face omnibus survey. Although it has its own biases, face to face research is generally regarded as the “gold standard” method of collection. The price of collecting data this way is commensurate to this status. We chose to weight the data both demographically and behaviourally – using both general questions about technology and specific questions regarding the mobile internet. This allowed us to calibrate our data to match the total UK population, for more accurate results.
Sample size: Online research is well suited to speaking to large amounts of people in a short space of time. We surveyed over 2,000 people in less than a fortnight. Due to the unique experiences of owning and using a mobile, a large sample size was crucial. It enables us to hold different variables constant and facilitate detailed comparisons across a range of criteria. Factors such as network, method of purchase, type of payment model, level of monthly spend, handset manufacturer, handset model and services used all factor into some one’s experience of mobile and it is important to understand how these factors differ by type of person.
Survey scope: The mobile internet space is a huge topic, but breadth of information had to be balanced with depth. Rather than superficially touch upon every area, we instead concentrated upon the key themes that emerged from the previous stages of research – the current environment, the key drivers and barriers to future consumption, and interest and attitudes to different kinds of services and media on the mobile. Future-facing information is important, but at this stage it can only be indicative. Full and accurate predictive behaviour can only emerge through detailed ethnographic user testing.
Analysing the data: There are many powerful techniques that give the ability to make additional inferences with the data. Multivariate analysis is an example of this. For instance, we were able to compare what people claimed their barriers to adopting the mobile internet were, with inferences as to what was actually the most important factors – derived from correlating specific factors with a likelihood to adopt. The results were strikingly different.
Creating the outputs: It is clear that the data collected from Brandheld is powerful, and relevant for many different applications. The outputs needed to reflect this flexibility. For instance, a couple of key figures illustrating the major themes would be most suitable for a topline executive summary. While an easy to read reference guide would benefit from simple bar charts highlighting comparative trends. However, since the anticipated uses are going to be so different and so unique to the particular person using the data, the best output is to simply make the raw data available. It can then be manipulated and filtered to answer even the most specific situations.
Yet the most powerful outputs are those that combine each of the phases of the research programme. Across the stages we have been able to develop hypotheses, iterate them through qualitative research, concept test them through ethnography and then statistically prove them through quantitative research. Combined, the research programme is greater than the sum of its parts.
Recent Comments