Moving surveys online for data collection and reporting
Online data collection has become the most common…Read more
Market research has adopted technology at varying speeds in its history, often at a different rate to other industries. Why is this? Is it important? CAPI (Computer assisted personal interviewing), CAWI (web interviewing), CATI (telephone interviewing), tabulations, reporting and dashboarding all pose different wants and needs. This article aims to look each type of software used in market research and discuss the importance of each type being online, offline or both – and, more importantly, where market research should be heading and what MRDC Software’s perspective is as a research software specialist.
Let me try to summarise market research as far as technology in one paragraph. As an industry, it led back in the 1970's by being proficient and making complex analysis and large-scale analysis possible. It adapted to the opportunities offered by CATI in the 1990's, but then advances started to fall behind other industries. Whilst technology roared ahead in most businesses from the late 1990's, market research moved forward more slowly. The PowerPoint presentation is still the end product of a serious slice of market research business. In the future, market research needs to be as convenient as the many apps most of use day in day out without much thought.
The arrival of online surveys meant a big change in research. It meant a lot of things, but predominantly it meant cheaper, faster, shorter surveys. This led to the demand for faster delivery of results. But, the end product has tended to remain as tabulations and PowerPoint presentations. Why is that? And, shouldn’t everything be online? Before we look at each type of software in a little more depth. Does it matter if software is online?
In my view, what goes on the ‘engine room’ doesn’t matter. What the client sees does matter. I don’t really care how my car is built, whether robots are used, whether someone puts parts together manually – all I care about is whether the car I drive serves me the way I want. The same is true about market research. So, to extend that argument, where the client ‘touches’ market research is at the end when they receive the results – that’s the bit where market research has been slow to move.
I would argue that market research as an industry has tended to make the ‘engine room’ bits more efficient. We can produce hundreds, thousands of tabulations quickly and we can prepare questionnaires efficiently, but these are things that the client largely doesn’t see. These things keep our costs down, but do not improve the customer’s customer journey if I can use a term that is overused in our industry. So, let’s look at each type of software in more depth.
At first sight, it doesn’t matter whether a CAPI questionnaire is prepared online or offline on a desktop or laptop computer. However, the way we work is changing. Without realising it, perhaps, we are getting used to being connected. This means that we can share files quickly or show our screen to a colleague or client online in seconds. Questionnaire development used to be a lonely task, but communication means that we need to be online, in my view. Trying to explain an issue to a client by email or by telephone takes far more time than allowing a client access to view themselves. It’s better if it is online.
The actual interviewing needs to be online and offline. Poor reception, black spots and unpredictable signal strength occurs over much of the world. The software collecting the data needs to work both online and offline by which I mean that if there is an inadequate internet connection the data is stored on a local device and uploaded, preferably automatically, when there is a connection. There are still several CAPI products that work either online or offline only – but not both.
For questionnaire preparation, the same is true for CATI surveys as it is for CAPI surveys. For interviewing, it is generally unimportant. Clients will see no benefit except for the fact that they cannot login and monitor project progress. I would also argue that being reliant on an internet connection could be a risk. If the connection drops for any reason, you could have 20 or more interviewers unable to work or with unfinished interviews. Of course, the same is true of a network, but at locations where internet speeds waver, a hardware network is usually preferable.
For questionnaire preparation, it is a must. Clients expect to be able to see their survey ‘in action’ and how well targets are being met. This is simply because we are all used to seeing what is going on when things are online, so for online surveys transparency is a must.
Producing tabulations or running statistical analysis is generally a hidden operation at most research agencies. I don’t think clients need to see the heavy number crunching that we regularly carry out on market research projects. Added to this is the fact that processing can be heavy and desktop processing is simply faster and more productive.
However, access to the data for quick counts or key figures is important. This can be approached in two ways – one is to use a system that is fully online, the other is to have a back end tool that allows clients or data users access to the data online or in some easily distributed way. The first method works well for many online surveys as online surveys tend to contain smaller amounts of data with shorter questionnaires. The second method is more appropriate for large scale projects or tracking studies.
I am still pained by the way PowerPoint presentations are heavily used to disseminate findings from research projects. I recall back in 1997 that I presented some data and made the session interactive by setting up something in Excel that could perform calculations and charts on request during the presentation. Yet, still much research is presented as static reports. I hope this changes.
Dashboards and interactive analytics are the way to go forward. This is not new technology, but it has been slow to be adopted by market research for two key reasons – firstly, the software has been expensive and, secondly, the tools are not really geared to the peculiarities of market research data – for example, multiple response questions, hiding data with low bases, significance tests etc. There is some movement on this at last and it is an area that MRDC is keen to pursue.
My assessment – and I’m trying to be fair-minded – is that we are doing quite well but must try to do better. Looking at our products in turn.
MRDCL for tabulations – This product is aimed at heavy number crunching so being online is not a priority, but we are improving our secondary analysis tools. Resolve, our new secondary analysis software tool, works as an Excel add-in allowing users to produce tables and charts from projects handled in MRDCL and we are looking to move this into Google sheets next year.
Snap for CAPI, CAWI, Tables, Reporting – Snap is essentially on an online tool that can be used offline, although at present developing a questionnaire is carried out offline. This will change during 2019 as the tool will be usable completely online making sharing of projects much easier. This is a very positive step in my view.
QPSMR CATI – QPSMR is still a desktop software product. Work is in progress to make it online, but as a CATI product its usage will still be better in most cases using a hardware network.
Iris – Iris is a game changer. It has made online dashboards quick and significantly less expensive to develop and opens new opportunities. As it has been developed specifically for market research it means that the industry can move forward. I hope its early impact on the market keeps driving it forward.
In 20 years from now, everything will be online. At this point, I think it is important to give clients access to their data where and when they need it – at every touchpoint. What goes on in the engine room doesn’t matter, but what the clients see and feel might be the difference between market research growing as an industry or being lost to other associated industries.