How to handle market research data analysis and tables
Like a lot of computer-based tasks, data analysis and tabulations from market research surveys can be easy in many cases, but exponentially more time-consuming when things become more complex. Indeed, on a number of occasions, I have seen companies take what I believe to be 20 times more man hours than necessary because they do not have the right expertise or the right tools to do the job. This guide takes a look at different types of data and survey analysis and looks at what software and skills are needed.
About this guide
This guide will discuss ten types of data analysis and market research tables and grade them as follows:
|Rating||Level of software needed||Level of staff needed|
|*||Any survey software, typically free or low cost software [we offer Resolve free for secondary analysis]||Anyone who understands market research|
|**||Most survey software [we offer Snap and QPSMR]||Anyone with a good market research grounding and familiar with software|
|***||Most professional survey analysis software [we offer Snap and QPSMR]||Anyone with a good market research grounding and familiar with software|
|****||Professional survey software or scripting language software [we offer MRDCL, Snap and QPSMR]||Good knowledge of market research software or data processing professional|
|*****||Scripting language software [we offer MRDCL]||Market research data processing expert|
Types of data analysis and tabulations
- Cross tabulations * to *****
Strictly, a cross tabulation is one question (or variable) as the rows of your table and another question (or variable) as the columns of your table. Any product that can do more than topline counts (or frequencies) will usually have this functionality. However, it can quickly get more complex than that. For example:
- Filters (selection of subsets of data) may need to be applied
- Figure and percentaging options may be needed (decimal places, vertical/horizontal percents)
- Ranking may be needed
- Table statistics may be needed (mean score, standard deviation, error variance, median etc.)
- Significance testing may be needed
- Tables may need to be crossed by more than one question or variable
- Outputs may be needed as text, PDF or Excel, for example.
The more features you need, the further up the scale from * to ****, you will go. Over and above, these types of consideration, the volume of tables you are producing and the volume of data you are processing may be an issue. If you are regularly handling bigger tasks, you are more likely to need *** to *****.
- Summary tables – summary of rating scales / mean scores *** to *****
Summary tables of mean scores, net preferences (NPS) summaries, rating scales summaries etc. are more likely to be possible the high rating of the product. The consideration here is ‘do you want outputs in an exact way’?
- Tables showing volumes ** to *****
Sometimes, you want to produce tables that are based on volumes rather than people giving answers in a survey. Tables of this type are often referred to as volumetric tables showing figures in terms of money spent, fuel consumed etc. rather than respondents. Most products at the ** end of the spectrum will handle this but as calculations become more complex, you will need to move to the **** level.
- Hierarchical data **** to *****
Not many products handle hierarchical data. By hierarchical data, I am referring to data where there is more than one level of data – for example, where you interview a doctor and then ask the doctor about his last 5 patients or a number of patients. Here, you have two levels of data – doctors and patients. You could even have a third level as you could record each drug that each patient received, for example. For such tasks, you need advanced software. Surveys covering eating occasions or purchase occasions are similar.
- Tracking study data *** to *****
Tracking studies where the data is consistent are relatively easy to handle. However, you will need to be able to handle or merge multiple files from period to period and probably show trends. This might be a case of lesser software can handle such tasks but with a more powerful product you are better placed in the long term.
- Tracking study data with questionnaire variations *** to *****
Analysis of these types of projects can be difficult to manage even though conceptually it might seem easy. Products at the top end of the range tend to handle these types of task using powerful data handling tools whereas products at the lower end tend to handle by editing and reformatting data. It’s an important question to ask potential suppliers if you have such projects.
- Data or tables requiring complex calculations or algorithms *** to *****
Again, power wins, but with power comes greater learning. If you have one complex variable that you need per project, it doesn’t really matter if it takes one hour rather 10 minutes because it is slow to define. However, if you have many complex calculations, algorithms and variables, more power becomes a necessity.
- Projects with target, factor or rim weighting ** to *****
More and more products now offer target and rim weighting. However, some don’t, so check if it is a must have.
- Projects where data from a database needs to merged with survey data *** to *****
Few products can do this or, if they can, it is achieved by complex data manipulation. If this is a one off task, it is not a problem. If you need to plug in an updated database every week, it needs to be readable in situ. Again, many mid-range products would not have this facility.
- Data that needs to be available in other software * to *****
This comes down to the tools available for exporting or transferring from one system to another. Some products only allow storage of data in their own proprietary format, but most products these days are able to produce data in a number of common formats such as Triple-S, which is an industry wide standard used in market research (see www.triple-s.org), Excel (xls or CSV), ASCII/text format as well as SPSS format.
Who should use the software?
The further up the scale you go, the more learning that is required to use a product. Of course, the trade-off is efficiency and an inadequate product may either mean that you cannot deliver what you want to or that it will take you many times more than some using a more advanced product. Further, complex projects, such as big tracking studies can become more and more complex as time goes on if they are not managed correctly using the right software.
Working with a range of clients, I have advised on a number of different strategies. One recent experience encapsulated how a flexible approach can be beneficial. A client had a large volume of fairly straightforward online surveys. However, one project was much more complex – it was a tracking study taking place in a number of countries with several questionnaire variations in each country.
When I sat down with the client and looked at the cost of using a complex product for all of their projects, we budgeted that software, training and staff costs would be significantly higher than having a lower cost product and subcontracting the complex project to another company. This meant that cheaper software, low training costs and cheaper staff more than compensated for the annual costs of outsourcing to experts – MRDC, in this case!
So, think flexible.
Where are MRDC’s products
If you need guidance on what you need, we are here to help. There’s no point spending too much money on an advanced product if you don’t need to, but at the same time, changing products costs time and money, so getting the right mix is important.