So you’ve decided that online research is the way forward. You are ready and raring to go – but first you need to build a sample. There are a number of firms that offer to fill your online research portal with a panel of participants. While it is certainly fast, this might not be the best approach to online sampling. An article from Jonathan Deitch of IPSOS recently pointed out some of the challenges panel providers need to overcome in the future.
Perhaps you are on a strict deadline, or perhaps you have a strong relationship with your panel provider. They can still be a huge benefit to your research, and much of the industry could not function without them. However, this guide is for those who want to build their own unique online market research sample. It can sound like a daunting task, but broken down into these six steps, the process becomes much more manageable.
Before you can begin to even think about recruiting online research participants, you must first define your research objectives. B2B International have published an extensive guide on research objectives and how to go about creating them. What we want to focus on is how these will affect your research sample. The most important consideration to take from this is what population you will be targeting.
Are you looking to introduce changes to your core product offering? If so you need to understand which groups this will affect and how you want them to be represented in your research. Customers, potential customers, marcomm audiences and stakeholders all need to be represented in a way which accurately reflects influence. If however, you are gathering consumer feedback on a new marketing campaign – then perhaps it is only the intended audience that is required.
There are a lot of online data collection tools. At FlexMR, we have sixteen spanning the full qual and quant spectrum, from online focus groups, to long term panels, diary studies and more. The main impact that your data collection method will have is on the size of the sample. A project comprising of 3-4 focus groups would be best suited to a sample of 35-45 participants. However, a complex research project such as a customer panel requires a larger sample (often in the region of 300+ participants). Be sure to know what type of research you are conducting, in order to build a large enough sample.
For research of any size or scale, the best place to start recruiting is your pre-existing networks. Make use of email lists, customer databases and any other existing connections you have built. People on these lists will be those most valuable to you.
The opinions of your customers will almost always be more relevant than those of a panel. Not only are they already familiar with your brand but will be motivated by a desire to improve the brand and experience, rather than just the financial incentive. To read more on what motivates research participants, take a look at Paul Hudson’s article on Engagement in the Online Era.
There has been a lot of buzz recently about social market research. Social media listening tools have driven this discussion and market researchers have been quick to adopt such processes. While we believe social listening can be a benefit to research, it is important to consider it within a broad framework of tools that tackles the research objective from every angle.
However, social elements can complement research through the entire process, from introducing community-based elements during the project, to driving participant recruitment. A subset of snowball sampling methodologies, social media recruitment leverages the personal connections of individuals to reach a wider potential audience. By combining this with your organisation’s own networks, it is possible to build a large (and representative) sample in a short space of time.
However, the obvious danger with building participant lists in this way is sample quality. There are a number of dangers to look out for, including speeders, professional research participants, no-shows and more. The best way to eliminate these is through carefully constructed screening questionnaires.
Screening questionnaires (or screeners for short) serve two purposes. The first is to ensure that your sample fits the profile you are looking for, the second is to eliminate participants that could degrade your data quality. Speeders are participants that complete research tasks as quickly as possible and do the bare minimum. Their responses are not always reflective of their own thoughts, often writing the first thing that comes to mind.
The easiest way to screen potential speeders is through red herring questions. These are questions which will ask participants to simply select a particular answer. While most will read the question correctly, speeders will unlikely even read the question before making a choice.
Professional research participants are more difficult to spot. By asking the right questions in your screener and not alluding to the responses that will lead to inclusion, most of these will be filtered naturally. Finally, no-shows are those at risk of not participating in the research. Some will not even complete the screening questionnaire, others may not finish it – either way the majority of these participants will be removed at this stage of the research.
Finally, once you have started your research project – you need to manage online participant sample quality. Over time, some participants will drop out, it is only natural. However, as more and more drop out this can have an overall negative impact on your research results. To ensure your results don’t suffer, you should be regularly monitoring active and inactive research participants. Run online recruitment throughout the length of the project to top up your sample.
By following these six simple steps, it is possible to create your own high quality panel of research participants. Of course, this is not immune to some of the wider issues associated with market research samples. However, by controlling the process, it is possible to create a sample specific and relevant to your business – built from those segments and groups most valuable to you. What are your tips for improving research sample quality? Let us know in the comments below.