Tips For Utilising MaxDiff in Market Research
MaxDiff can help market researchers generate robust insights where a plethora of outputs are up for grabs. In part two of her blog, CMR Research Manager Sarah Mackinnon shares her top tips to help you maximize MaxDiff in your own study.
1. Start with qual to define your bucket list
In CMR’s experience, the most successful MaxDiff studies build on in-depth qualitative research. This is particularly important when it comes to NPD (new product development) – there’s no point spending your research budget on getting 500 patients to prioritise a list of ‘needs’ that don’t exist in the first place. Also, you don’t want to run the risk that some of the most crucial market needs may be absent from your list because they still remain unidentified.
Learn more about MaxDiff’s pros and cons in Part 1 of Sarah’s Blog: Why Choose MaxDiff?
At Creative Medical Research, we recommend a ‘bottom-up’ approach. Ideally, start with ethnography to identify unmet needs through non-participatory observation of target users. Ethnographic insights can then be taken forward for discussion in qualitative interviews, to add depth and detail. The rich qualitative data will elucidate the breadth and depth of market needs, which can then be quantitatively and robustly prioritised in your MaxDiff study.
2. To avoid priming bias, MaxDiff exercises should come as early as possible in your survey
Frequently, our clients seek to address multiple research objectives within the course of one survey. For example, in an NPD study, objectives may include gaining an understanding of current practices and their perceived advantages and disadvantages, as well as assessing needs for future products.
Logically, you might think such a survey should follow the order in the sentence above for best question flow. However, depending on what questions are asked in the early part of the survey, and of course how they are asked, this can affect the results you get for later questions. As a crude example, if you asked physicians a series of questions regarding safety issues with current devices, then presented a range of safety and performance benefits in a MaxDiff task, you’d surely see a stronger bias towards the safety benefits in the results because the physicians have been primed to focus on safety.
Of course, if the questions you ask before a MaxDiff are well balanced, equally covering the issues tackled by the list to be tested, later placement shouldn’t raise an issue, and indeed, survey flow may be improved. However, if you have any concerns, putting the MaxDiff exercise first up can alleviate potential headaches in data interpretation later down the line
3. Use 'on the fly' in-survey analysis to feed through results to follow up questions and contextualise the findings
As noted previously, one thing MaxDiff in itself can’t do is tell you the absolute importance of your items – only how important they are relative to each other. It’s therefore possible, though less likely if your study is built on a solid foundation of qualitative research, that even the top needs on your list may be relatively unimportant.
However, there are easy ways to get around this issue and fully contextualise your MaxDiff data. MaxDiff surveys can be programmed to prioritise individual respondent’s needs ‘on the fly’, allowing you to pipe their most (or least) important items into follow up questions that may assess.
So for example: priority to address these issues relative to others they may be facing in their wider practice; emotional response to potential product features; the extent to which favoured messages are felt to align with brand image, or open ended questions to probe the ‘why’ behind their choices.
4. Make the most of individual level scores with segmentation
As MaxDiff calculates individual scores for each respondent for all list items, results can readily be used to segment the marketplace; either demographically (at the simplest level), or using complimentary needs-based segmentation techniques such as Latent Class Analysis.
Segmentation allows us to move beyond simply understanding what the majority wants, enabling us to answer more complex research questions, such as ‘how can we best optimise our product portfolio in order to meet the differential needs of a wide potential customer base?’, or ‘how should we target our sales approach to each market, in order to maximise success on launch?’.
As such, if you really want to get the most out of your MaxDiff study, it’s usually worth the extra investment to include a Latent Class Analysis, or at the very least, budget for 5 minutes extra survey time to build a full demographic profile for respondents.