Context influences decisions. We should embrace it.

You sometimes hear in market research that people want objective data. Meaning data, which is not tarnished by bias, the context of the question or how the question was asked. This is often motivated by the desire to understand how customers perceive the actual features of a product. Allowing them to (supposedly) focus on assessing the features themselves without being distracted by external influences.  

This is an understandable wish, but it is also an impossible one. Rather than removing context to avoid being distracted when assessing the features, we should embrace context with open arms. By re-creating and assessing contextual factors we can deliver more realistic and helpful results.  

Here's why:

1. Context matters.

Market research can help us understand not only customer perception but also the specific context in which the data is collected. If you remove this context within a survey, you are removing key decision factors which could be influencing decision-making.

Studies conducted within behavioural science demonstrate that even seemingly irrelevant factors can have an influence on our decision-making. For instance, people are more prone to make optimistic stock investment decisions when the weather is sunny (Hirshleifer, D. and Shumway, T, 2003). Judges have been found to give more lenient sentences after a lunch break than when tired and hungry later in the day (Danziger, S., Levav, J. and Avnaim-Pesso, L. , 2011). Slower classical music has been found to increase spending within restaurants (North, A.C., Shilcock, A. and Hargreaves, D.J, 2003).

Research environments are not exempt from contextual influences.  Factors such as response time, information formatting, physical/social environment, incentive and mode of delivery all have the potential to impact customer responses.

The repercussions of removing these contextual factors from research design are clear; remove the context and you risk generating research results which bear little resemblance to how a product evaluation would pan out within a real buying scenario.

Of course, it may not be possible to recreate a real buying scenario entirely, but getting closer to the buying scenario gives you a better chance of more accurate results.

2. We can measure context.

If a certain way of communicating or presenting a product is biasing results, then let's test how it ‘biases’ results. All we need to do is manipulate the context the product is shown. If it's contributing to why a customer is selecting a product, then it's worth testing. Let’s look at one example from Spotify to think about how we might test context.

Spotify Playlists

Spotify were interested in evaluating the playlists they produce for listeners. They had trialled both “Throwback Thursdays” and “Feel Good Fridays” to relatively moderate results. However, the introduction of a third playlist option “Discover Weekly” on a Monday received a markedly better reception. Analysts were looking at data such as skips, saves, dwell time, and appeal scores, to understand why it had been better received.

One element which hadn’t been tested was the day on which the playlist was trialled. In this case, trialling the playlist on a Monday provided the perfect context for increasing use of the playlist. Monday is the day of ‘fresh starts’ and so matched well with the playlist theme ‘Discover Weekly’ (Page,W. & Bursztyn, M.,2019).

The impact of the day the playlist is released cannot be given to us directly by Spotify listeners, but it can be inferred by conducting two tests on the appeal of a Spotify playlist where the only difference for the second test is the day the playlist is tested.

This manipulates the context to give us the specific effect context has on engagement in the playlist. By using this experimental technique, we open up a new way of measuring a product which goes beyond what can be self-reported.

3. We can change the context.

If you can improve the uptake of a product then it doesn't matter whether you do this through changing the product’s benefits or the context it is shown in. They both contribute to the same target behaviour.

It is when we focus on the target behaviour in product and service evaluation, that we unleash the power of context and discover a new set of solutions.

Let’s apply how we might do this to our Spotify playlist example. If you are a traditional market researcher, you may look at evaluating this product by saying “let’s understand what Spotify listeners like or dislike about this playlist”. Though this approach will give insight into listener perception, it is by focusing too narrowly on appeal of the product itself that we miss out on a whole host of potential improvements which could increase product engagement.  

A better question to ask here is; “what changes can we make which will increase engagement with the Spotify playlist?”. This could be a change with the playlist itself or indeed with any context surrounding how the playlist is delivered.

Though he’s not a Market Researcher or Behavioural Scientist, the former US President Harry Truman encapsulated this sentiment by saying “It is amazing what you can accomplish if you do not care who gets the credit.”

The focus of product evaluation should be what behaviour we are looking to change and how we can change it. To do this, any part of the product offering can take credit for what changes behaviour whether it be the product ingredients, the way its packaged or the messaging.

This is the strange thing about removing context in Market Research testing, you are removing the testing of potential solutions which can be offered to a client, due to trying to focus too narrowly on one aspect of the product offering.

You see this every time a blind taste test is conducted. By trying to isolate the taste of the product, you are removing a potential idea for a branding change which could increase interest in the product. The mindset shift here is in understanding that, at the end of the day, it doesn’t matter whether interest in the product comes from the product features or the context in which the product is shown in. Both contribute to the target behaviour. Both are worth testing. Both can be altered to enhance the product.

So, let's remember...

Research without context is not objective research, it is limited research. The best research will be conducted only when all product or service solutions are on the table. When we remove context from testing altogether, we remove potential solutions which could be offered to clients. We must then recreate and measure context in research so we can apply it to solutions.

I would suggest that the next time you conduct or commission a research project, you remember to consider how you can broaden your testing by considering context. You never know when seemingly irrelevant context can drive customer behaviour.