Social listening and analysing online reviews is extremely insightful when it comes to understanding your audience and their relationship with your brand. But what if you have a specific line of questioning in mind?
Perhaps you want customer feedback immediately after a product launch, and there isn’t enough data online yet. Or maybe you’d like to explore what your customers are saying in more depth.
Surveys are a popular way to garner specific information from your audience, but unless you know how to analyse survey data effectively, the process can be time-consuming and may not give you the insights you need.
Today we’re offering you tips and advice to help you get the most out of your survey data. And it all starts with the survey itself…
1. Create a useful survey
The emphasis here is useful. Before you start preparing your questions, first understand the purpose of your survey. For example, is it to improve your marketing strategies, establish the direction of your brand, or product innovation?
You want to keep your questions to an absolute minimum to keep respondents engaged through to the completion of the survey.
According to research conducted by OpinionLab, 80% of customers abandon a survey before completing it, and 52% said they’d spend no longer than 3 minutes filling out a feedback form.
Survey fatigue can also decrease the quality and reliability of the answers provided.
With this in mind, try to combine two or more questions. One way to do this is by asking open-ended questions (questions that require more than just a yes/no response). Historically, surveyors may have avoided open-ended questions to help categorise and analyse their data. However, thanks to advancements in text analysis technology, qualitative data is easier than ever to analyse.
2. Ensure you have a good sample size
Once you start slicing your data into different segments and subgroups, your sample sizes will become smaller and smaller. Before you begin analysing, make sure you have enough data to play around with. The larger your sample size, the better, and with the help of specially designed analytical tools (see more below) you needn’t be overwhelmed with big datasets.
SurveyMonkey has created a survey size calculator to help you estimate the appropriate sample size.
3. How to analyse quantitative data
Start by aggregating your quantitative results. This means totting up the total number of people who answered with each response. For example, if 100 people were asked what their favourite fruit is, your aggregated results would look like this:
Convert raw data into percentages to help you make sense of the proportion of respondents that gave each answer.
Now you want to delve in and get insights by comparing two data sets. Taking the earlier example, let’s say you want to compare people’s favourite fruit by age group
|Apple||Melon||Banana||Orange||Apple %||Melon %||Banana %||Orange %|
By cross-tabulating, you can identify patterns and see if there are any correlations. You can easily do this in an excel spreadsheet, though it’s not the most time-efficient option, and some tools can do the work for you.
4. How to analyse qualitative data
Now it’s time to delve deeper and discover the how and why behind your quantitative data. Until recently, analysing qualitative data has been arduous and time-consuming, and text analysis tools were unreliable. But machine learning and AI technology have come on leaps and bounds in recent years, meaning that it’s easier than ever to analyse vast quantities of qualitative data in detail.
The Symanto Insights Platform is a text analysis tool that enables you to get both a broad overview of what people are saying, and delve deep to explore survey answers on a granular level.
As well as categorising answers by topic, and sentiment, Symanto also uses natural language processing (NLP) to unearth the personality traits and communication styles of respondents. This allows you to get further insight into your customer base based on what they say and how they say it.
The Symanto Insights Platform automatically aggregates and cross-tabulates your qualitative data so you can get relevant insights within a matter of minutes. Furthermore, the platform is already connected to Typeform and SurveyMonkey, making it particularly easy to analyse your survey data if using either of these tools.
5. Be critical
You’ve spent time and energy creating and analysing your survey, you are invested in getting some juicy insights. But beware of analysis bias – searching for data that reinforces a theory while ignoring contradictory data and/or overstating the significance of the data.
Cast a critical eye over your data and your insights, particularly if your sample size is on the small side or is unrepresentative.
Consider correlation vs causation. Looking at the example of favourite fruits in the table above, you could conclude that people favour apples as they age. This infers causation (ageing causes apple preference) when the cause may be unrelated – perhaps the boomer generation has a lifelong preference for apples that won’t be reflected in subsequent generations. Nonetheless, there is a correlation that may be worth investigating further.
6. Report on insights, not just on data
When presenting your findings to stakeholders, you want to lead with insights, not just data. Percentages can help you illustrate your point, but don’t overegg it. Stakeholders may want to know the implications of your data, they may want to know what potential action could be taken, but they don’t want to know the ins and outs of your research.
For more information on improving your insights, read our blog “How AI Solutions Can Enrich The Quality Of Your Data.”