Despite decades of poor results we still regularly see organisations carrying out qualitative research that does not appear to help their cause – and in the worst cases leads to initiatives that are bad investments.
Here’s an example which illustrates where this research is going wrong. What’s the problem with the following question?
‘How long would you like to wait at A&E?’
There is a good deal wrong with this question, but the key problem is the answer that the respondent will give you. Unsurprisingly most people answered this question by specifying a time-frame of around 5 minutes.
As an answer, this gives next to no information with which to make decisions and choices that could improve the patient’s experience. It’s a wasted question.
Unfortunately this was a bona fide question from a very high profile and comparatively expensive piece of research – proving that this type of loose thinking is far from a simple rookie mistake. Let’s go back to the fundamental purpose of customer research. We do it to gather intelligence and insight about our customers, which we hope to use to guide decisions and choices – improving, for example, our brand, the propositions we sell, the markets we compete in or wish to enter, and the service we give to the customer. At a general level there are two main research approaches: qualitative (qual) and quantitative (quant). This article deals with the qual side. Quant work usually involves large data sets, and is almost always gathered electronically and where we can have some statistical certainty about the findings. Qual is less likely to be statistically relevant but has other benefits if used in the right way, particularly when there is a lack of good quality quant data and we are looking to find at least some direction. The options for qual research can range from one-to-one interviews through to focus groups and surveys of various sizes.
Unfortunately, there often seems to be two mind-sets in business about qual research: the first says it is meaningless, as the smaller scale has no statistical robustness; while the second is willing to use a single statement from a single customer at one focus group to justify major change. In our view neither of these is right; using qual and quant research together is the best way to achieve success.
We often advise our clients to blend both elements of research, so that the quant research (what, how much, how many?) is backed up or ‘coloured’ by the qual research – so that emotion and reasoning can be understood in the context of a customer’s decision. This allows us to hear the voice of the customer but have some certainty over the choices they make.
Let’s return to our A&E example above: ‘How long would you like to wait at A&E?’ What we are really looking for here is information to help improve patient satisfaction with A&E waiting times. Therefore, a combination of the quantitative question ‘How long a wait do you consider reasonable in a busy A&E?’, where the respondent is asked to choose one option from a (carefully chosen) range of answers, and the qualitative question ‘What do you find most frustrating about long A&E waiting times?’ will reap many more actionable insights. Responses to the first question will give an appropriate target time to aim for while the responses to the second will help you identify strategies to improve patient satisfaction at those times when a long wait is unavoidable.
Recent research carried out by a major supermarket exemplifies why a qual approach is important. As customers were entering the store, researchers asked them whether they bought Fairtrade bananas, and then looked at the sales of Fairtrade bananas at the till. Over 30% of customers said that they would buy Fairtrade bananas prior to making their purchases, but only 4% actually did. So does this mean the supermarket should drop the Fairtrade bananas?
Maybe, maybe not. The customer’s aspiration may well have been to buy Fairtrade bananas – and this is important for future decision-making and ranging, but the actual numbers were much lower. Had the supermarket stocked their shelves with 30% Fairtrade bananas, then they may well have lost money. This is where qual research can help – by identifying why the customers didn’t purchase the Fairtrade bananas: was it simply offers, positioning and base-level pricing, or were there other factors at work – for example self-image vs reality. The scale of difference here is so significant that we need to take notice and ensure that we set up any qual research with quant input to ensure balance.
How can PA help?
We recently worked with a membership body to help them carry out research. We provided:
- The right focus – research agencies are often one step removed from the underlying problem. It’s one thing to recognise the issue that needs to be addressed, it’s quite another to fully understand your client’s commercial, political, and market situation to ensure that you do not ask ultimately useless questions.
- The right design – firstly, it is important to ensure that the questions asked in a survey are specific enough, and that the respondent is qualified to give a response. This is not to say that we should ‘lead the witness’, but questions do need to guide the respondent to a degree that allows us to understand their preference, eg for pricing in a certain range. Secondly, we must consider that the respondent is often not qualified to give an answer. Going back to our A&E example, were the survey designers right to ask the person waiting or should they have thought harder about the design before asking the question? Certainly, Apple would argue that, if left purely to the customer, the iPod may never have been designed.
- The right analysis – a specific piece of research should be able to link up with other research taking place in the business, in order to provide a more accurate picture and enable the company to compare the insights.
So how can we improve this situation? Our key piece of advice is to start with the end in mind and work backwards. Having clarity over the outputs that you are looking for from the research is crucial and should guide all decisions; including your audience, what type of research, and ultimately the questions you ask. Going back to our A&E example, we know that waiting times are a key issue – but so are many other matters. We also know that waiting times are only ever likely to be in a range, so using options for questions within certain parameters is more likely to provide useful outputs. It is important to look at a question, its context, and the method of research and consider how you will use the answers to improve business. If it can pass these acid tests, you are likely to get significant value from the investment.
If you would like to find out more about our customer research services and insights, please contact us now.