Data analysis and reporting brings all your previous efforts to the final act of the research process, and it’s where you will need to skill-up and practise to achieve excellent research.
There’s a lot to consider in the final stage of the research process. Firstly, go back to the brief. Ask yourself, what problem was the business trying to solve? What are the strategic decisions that will be made off the back of the research? What specific groups or segments will need to be considered?
Secondly, who is the audience? What are their roles? What knowledge will they have? How much time will they have?
Thirdly, how will you be sharing your findings? Will you be presenting? What level of detail is required? Will there be different analyses and reports for different groups or stakeholders?
Next, create an analysis plan so that you avoid analysis paralysis. Think about what you will focus on. Map out the analysis before you even look at the data. Use the questionnaire as an index to the data but don’t just follow the flow of the questionnaire – you need to tell a story.
There is a lot to consider including filtering questions, bases, weighting, creating statistics, analysis banners and statistical testing to discover real differences. Most of the readily-available survey packages offer some sort of basic analysis tools, but be careful with your analysis and check it thoroughly to ensure it is correct, and before you start drawing conclusions and writing a report. You may need to invest in more sophisticated statistical software to achieve all your analysis requirements.
Will you weight the data using known population parameters such as ABS census data? The benefit of weighting is that it not only smooths your data and brings everything back to its true proportion in the population; you can also refer to your results in terms of number of people (thousands) in addition to percentages. Weighting is a sophisticated process and requires specialised knowledge and software to execute properly, so consider carefully if and how you will weight your data.
Then start building a banner for cross tabulations. What are all the discriminators you might want to look at across the sample such as age, gender and area? Consider other variables such as attitudes or category/product usage.
It’s also important to consider the different types of variables and then matching them with the right analysis.
- Nominal data such as gender (and other “labels”) need to be analysed as a simple incidence or distribution.
- Ordinal data (ranking or natural order) such as agree/disagree scales might require incidence, range, “top 2 box” and median scores (not mean).
- Interval data e.g. Net Promoter Score 0–10 recommendation ratings, might include the mean, variance and standard deviation.
- Ratio data (e.g. price paid) would include averages, variance, standard deviations and ratios.
Familiarising yourself with the various type of data will be useful in getting this right and there are many resources available online to help you understand these.
Think about what statistical procedures could add value to the data, including:
- Regression analysis to understand what drives a particular variable such as satisfaction.
- Factor analysis for reducing data to understand themes.
- Cluster analysis to segment consumers.
- Perceptual mapping to understand brand positioning and whether you want to overlay other data sources such as sales data.
This is perhaps the most specialised skill required in the survey process. If you are not comfortable with how to process the data, then perhaps think about investing a small sum into professional help. There are plenty of data bureaus that can help, and D&M Research is an excellent choice.
Next, you will want to consider how you will turn these findings into a report or presentation. This is where you will tell the story of the data and highlight not only new insights but the evidence in support of or against different decision directions.
There is no set way to successfully write a research report and different practitioners have different approaches that are refined after many years of experience.
Here are some universal guidelines that can help you develop your own approach.
You need to develop a storyline as you are effectively turning numbers into words.
Think of it as a movie with different scenes; use post-it notes to develop these scenes (which can be topics or the findings you want to highlight) and arrange them in an order that creates a storyline and flow. Don’t worry if you change it later; this will most definitely happen as it’s an iterative process, so don’t be too precious about the order. The main thing is that you have a process for generating the story.
Here is a good starting point for an overall flow
- Table of Contents
- Background & Objectives
- Key Findings
- Detailed Findings
- Conclusions and Recommendations
Create charts and graphs that support the story rather than the other way around
Spend time thinking about this; don’t just run off and generate a mountain of charts that don’t achieve anything other than creating a door stop! Think also about the suitability of chart types e.g. pie charts for single-response questions that add up to 100%; bar charts for multiple-response questions; line charts for tracking data; brand maps for positioning data, etc.
Try to get the right balance between detail and summary findings
This is a tricky step to get right and will depend very much on who the audience is. Some clients (or people) prefer a lot of detail; others (usually more senior) just want to know succinctly what they need to know and what your recommendations are.
Production values are also important
Our own research shows that if a report looks professional, then it is also generally seen as accurate and believable. This includes ensuring that you do not have typos and inaccurate data. Everything needs to be proofed and checked, but beyond that you need to use well-designed templates and style guides to ensure that your work shines. Be consistent in this style and limit the number of colours and fonts used for headings and main body text.
The quality and value of the content is crucial to creating an excellent report
The report needs to encompass the objectives and marketing issues and these should be reframed from the briefing document. The story you create from the findings should address these and build on them, creating a compelling presentation that leads to your conclusions and recommendations. If you’ve done a good job, these conclusions should seem logical to the audience and your recommendations should also then flow as logical choices.
Ensure that the recommendations you make fit with the decision options the client has; don’t recommend directions that the client would never consider for cost, capability or brand fit reasons, have already been flagged as off-the-table or are simply too radical to be considered. Verbal consultation with your client or stakeholder is a good way to test the waters with any recommendations before writing them down or presenting them.
Finally, be proud of your work
Before you are ready to say, “it’s finished”, ask yourself, “is this my best work?” If it’s not, then there’s more to do. If it is, then be proud of the efforts you have made to get to this point and relish the opportunity to present your work and share your new findings and insights.